Thread: Re: [Semanticscuttle-devel] Chat with Christian Weiske
Brought to you by:
cweiske
Menu
▾
▴
semanticscuttle-devel
|
From: Christian W. <cw...@cw...> - 2011-05-12 12:39:41
|
Hi Mark, > > 3:36 AM Christian: do you have an idea how to solve the session > > problem? I'd probably use a static class variable > > we need to see if logging in/logging off and all the unit tests > > run afterwards > Would you be kind enough to catch me up on the actual session problem? 1. The private key authenticates the user on certain areas of SemanticScuttle - currently only on the RSS feeds. No other area. 2. To be able to use the service methods required for private RSS feeds, we need to register the user identified by the private key as the current user. 3. The current user is stored in the session 4. To restrict privatekey access to the RSS feeds only, we need to unregister the current user after the feed has been generated. 5. If the script crashes during RSS generation, the user does not get unregistered/unset. This means that the user will be still logged in and can access all areas although he originally only had the private key. So we need to decouple the current user from the session storage. -- Regards/Mit freundlichen Grüßen Christian Weiske -= Geeking around in the name of science since 1982 =- |
|
From: Christian W. <cw...@cw...> - 2011-05-12 18:09:51
Attachments:
signature.asc
|
Hello Mark, > > 1. The private key authenticates the user on certain areas of > > SemanticScuttle - currently only on the RSS feeds. No other area. > > 2. To be able to use the service methods required for private RSS > > feeds, we need to register the user identified by the private key > > as the current user. > > 3. The current user is stored in the session > > 4. To restrict privatekey access to the RSS feeds only, we need to > > unregister the current user after the feed has been generated. > > 5. If the script crashes during RSS generation, the user does not > > get unregistered/unset. This means that the user will be still > > logged in and can access all areas although he originally only had > > the private key. > > > > So we need to decouple the current user from the session storage. > Gotcha. About to walk into a meeting. I think it can be done, and > hopefully with little pain. Will catch up later. I fixed that in the master branch; just merge it - it should apply without problems. -- Regards/Mit freundlichen Grüßen Christian Weiske -=≡ Geeking around in the name of science since 1982 ≡=- |
|
From: Mark P. <mpe...@gm...> - 2011-05-13 03:15:38
|
Whew, ok, I either passed or failed. Please check my work to make sure we're on the same page. If I failed, don't be too harsh on me. Everything seems to be in order with the rss feed. On Thu, May 12, 2011 at 2:09 PM, Christian Weiske <cw...@cw...>wrote: > Hello Mark, > > > > > 1. The private key authenticates the user on certain areas of > > > SemanticScuttle - currently only on the RSS feeds. No other area. > > > 2. To be able to use the service methods required for private RSS > > > feeds, we need to register the user identified by the private key > > > as the current user. > > > 3. The current user is stored in the session > > > 4. To restrict privatekey access to the RSS feeds only, we need to > > > unregister the current user after the feed has been generated. > > > 5. If the script crashes during RSS generation, the user does not > > > get unregistered/unset. This means that the user will be still > > > logged in and can access all areas although he originally only had > > > the private key. > > > > > > So we need to decouple the current user from the session storage. > > Gotcha. About to walk into a meeting. I think it can be done, and > > hopefully with little pain. Will catch up later. > > I fixed that in the master branch; just merge it - it should apply > without problems. > > > -- > Regards/Mit freundlichen Grüßen > Christian Weiske > > -=≡ Geeking around in the name of science since 1982 ≡=- > > > ------------------------------------------------------------------------------ > Achieve unprecedented app performance and reliability > What every C/C++ and Fortran developer should know. > Learn how Intel has extended the reach of its next-generation tools > to help boost performance applications - inlcuding clusters. > http://p.sf.net/sfu/intel-dev2devmay > _______________________________________________ > Semanticscuttle-devel mailing list > Sem...@li... > https://lists.sourceforge.net/lists/listinfo/semanticscuttle-devel > > |
|
From: Christian W. <cw...@cw...> - 2011-05-13 05:31:28
Attachments:
signature.asc
|
Hello Mark, > Whew, ok, I either passed or failed. Please check my work to make > sure we're on the same page. If I failed, don't be too harsh on me. > Everything seems to be in order with the rss feed. Could you please look through the last QA list I sent? I just began testing and many items are still open. I also saw that the opensearch unit test does not work because of a leading empty line. Did it really work for you? And I'd like to have another feature, that is private key support for tag feeds (add it to the meta there, too). About your latest changes: www/rss.php: You have 2 times the same privatekey code, which is unnecessary in my eyes. Service/User.php::loginprivatekey(): you set the session variable instead of using setcurrentuserid. I've got to say that I have problems comparing the privatekey branch with master because of the many unrelated changes in there. The diff is a whopping 214kiB. Since I don't want to throw that changes away and lose your work, I'll gradually merge whitespace changes from your branch into master. Maybe it would be helpful to start a new branch with a fresh clone of the master branch and cherry-pick the relevant changes from the current privatekey branch. This way we'd have a clean pk branch and overview about all changes. -- Regards/Mit freundlichen Grüßen Christian Weiske -=≡ Geeking around in the name of science since 1982 ≡=- |
|
From: Mark P. <mpe...@gm...> - 2011-05-13 12:51:05
|
Here's the last list that I can find: > + Privatekey works fine for the user's own bookmarks > - Does not work with all bookmarks (plain rss.php?privatekey) - I'm > likely to subscribe to the all bookmarks feed in a private SC setup > with a several users. > + Enabling and disabling the private key works fine > + Error page when accessing feed with wrong private key > + Error page when accessing feed with disabled private key > - Generating new key throws away form changes. This isn't that nice, > but I'll accept it for now and we can fix that when the quickform > branch is finished. > - 1) BookmarkTest::testPrivateBookmarks > Undefined offset: 1 > /home/cweiske/Dev/semanticscuttle/cwdev/tests/BookmarkTest.php:1387 If this is the same list you are referring to, I thought we went through all of them. Will confess that I did see some unit test errors and will work on them. I also had a lot of conflicts when merging master, which I thought was normal. No excuses on my part. Here's a new working to-do list. Please add as you see fit: Before continuing I'll create a new branch from master (calling it primarykey2) and pull in all relevant changes. 1) Resolve unit test errors 2) Refine code in rss.php 3) Update User.php::loginPrivateKey() to follow updated user session configuration 4) Add private key support to tag feeds Cheers, Mark On Fri, May 13, 2011 at 12:56 AM, Christian Weiske <cw...@cw...>wrote: > Hello Mark, > > > > Whew, ok, I either passed or failed. Please check my work to make > > sure we're on the same page. If I failed, don't be too harsh on me. > > Everything seems to be in order with the rss feed. > > Could you please look through the last QA list I sent? I just began > testing and many items are still open. > > I also saw that the opensearch unit test does not work because of a > leading empty line. Did it really work for you? > > And I'd like to have another feature, that is private key support for > tag feeds (add it to the meta there, too). > > About your latest changes: > > www/rss.php: You have 2 times the same privatekey code, which is > unnecessary in my eyes. > Service/User.php::loginprivatekey(): you set the session variable > instead of using setcurrentuserid. > > > I've got to say that I have problems comparing the privatekey branch > with master because of the many unrelated changes in there. The diff is > a whopping 214kiB. Since I don't want to throw that changes away and > lose your work, I'll gradually merge whitespace changes from your branch > into master. > > Maybe it would be helpful to start a new branch with a fresh clone of > the master branch and cherry-pick the relevant changes from the current > privatekey branch. This way we'd have a clean pk branch and overview > about all changes. > > -- > Regards/Mit freundlichen Grüßen > Christian Weiske > > -=≡ Geeking around in the name of science since 1982 ≡=- > > > ------------------------------------------------------------------------------ > Achieve unprecedented app performance and reliability > What every C/C++ and Fortran developer should know. > Learn how Intel has extended the reach of its next-generation tools > to help boost performance applications - inlcuding clusters. > http://p.sf.net/sfu/intel-dev2devmay > _______________________________________________ > Semanticscuttle-devel mailing list > Sem...@li... > https://lists.sourceforge.net/lists/listinfo/semanticscuttle-devel > > |
|
From: Mark P. <mpe...@gm...> - 2011-05-13 21:59:32
|
Status: I've created the new branch from main, and just finished migrating over all the privatekey changes. About to head out of town On Fri, May 13, 2011 at 8:50 AM, Mark Pemberton <mpe...@gm...>wrote: > Here's the last list that I can find: > > + Privatekey works fine for the user's own bookmarks > > - Does not work with all bookmarks (plain rss.php?privatekey) - I'm > > likely to subscribe to the all bookmarks feed in a private SC setup > > with a several users. > > + Enabling and disabling the private key works fine > > + Error page when accessing feed with wrong private key > > + Error page when accessing feed with disabled private key > > - Generating new key throws away form changes. This isn't that nice, > > but I'll accept it for now and we can fix that when the quickform > > branch is finished. > > - 1) BookmarkTest::testPrivateBookmarks > > Undefined offset: 1 > > /home/cweiske/Dev/semanticscuttle/cwdev/tests/BookmarkTest.php:1387 > If this is the same list you are referring to, I thought we went through > all of them. Will confess that I did see some unit test errors and will > work on them. I also had a lot of conflicts when merging master, which I > thought was normal. > > No excuses on my part. Here's a new working to-do list. Please add as you > see fit: > > Before continuing I'll create a new branch from master (calling it > primarykey2) and pull in all relevant changes. > > 1) Resolve unit test errors > 2) Refine code in rss.php > 3) Update User.php::loginPrivateKey() to follow updated user session > configuration > 4) Add private key support to tag feeds > > Cheers, > Mark > > > On Fri, May 13, 2011 at 12:56 AM, Christian Weiske <cw...@cw...>wrote: > >> Hello Mark, >> >> >> > Whew, ok, I either passed or failed. Please check my work to make >> > sure we're on the same page. If I failed, don't be too harsh on me. >> > Everything seems to be in order with the rss feed. >> >> Could you please look through the last QA list I sent? I just began >> testing and many items are still open. >> >> I also saw that the opensearch unit test does not work because of a >> leading empty line. Did it really work for you? >> >> And I'd like to have another feature, that is private key support for >> tag feeds (add it to the meta there, too). >> >> About your latest changes: >> >> www/rss.php: You have 2 times the same privatekey code, which is >> unnecessary in my eyes. >> Service/User.php::loginprivatekey(): you set the session variable >> instead of using setcurrentuserid. >> >> >> I've got to say that I have problems comparing the privatekey branch >> with master because of the many unrelated changes in there. The diff is >> a whopping 214kiB. Since I don't want to throw that changes away and >> lose your work, I'll gradually merge whitespace changes from your branch >> into master. >> >> Maybe it would be helpful to start a new branch with a fresh clone of >> the master branch and cherry-pick the relevant changes from the current >> privatekey branch. This way we'd have a clean pk branch and overview >> about all changes. >> >> -- >> Regards/Mit freundlichen Grüßen >> Christian Weiske >> >> -=≡ Geeking around in the name of science since 1982 ≡=- >> >> >> ------------------------------------------------------------------------------ >> Achieve unprecedented app performance and reliability >> What every C/C++ and Fortran developer should know. >> Learn how Intel has extended the reach of its next-generation tools >> to help boost performance applications - inlcuding clusters. >> http://p.sf.net/sfu/intel-dev2devmay >> _______________________________________________ >> Semanticscuttle-devel mailing list >> Sem...@li... >> https://lists.sourceforge.net/lists/listinfo/semanticscuttle-devel >> >> > |
|
From: Christian W. <cw...@cw...> - 2011-05-14 08:06:13
Attachments:
signature.asc
|
Hello Mark, > If this is the same list you are referring to, I thought we went > through all of them. I *thought* I mentioned that the create-testbookmarks script had a limit of 10k instead of 10 bookmarks in it, and that's what bothered me. But I looked through the mails and didn't find it. Consider my rant baseless. -- Regards/Mit freundlichen Grüßen Christian Weiske -=≡ Geeking around in the name of science since 1982 ≡=- |
|
From: Mark P. <mpe...@gm...> - 2011-05-17 05:18:19
|
Unit errors: 1a) One in Bookmark2TagTest::testGetPopularTagsDays appears to exist in master, so I can't claim that. I spent some time trying to resolve it, but was unable to. 1b) Another unresolved issue is in Tag2TagTest::testSearchThroughLinkedTags which is a result of a change I did in getBookmarks(). I'll go ahead and apologize now if I'm wrong in my understanding of how the tag2tag's work. The change I made seems to make sense, and seemed to produce the results that I was expecting. However, when I saw this test fail, red flags started raising and now my change is in doubt. So either the update to getBookmarks() is all wrong and all of it needs to be rejected, or it is right and the test needs to be updated. Regardless if I'm right or wrong, I'd love to use this opportunity to properly document this area of work to help eliminate confusion in the future. 1c) The rest of the unit errors are in the API and it appears to be my environment that is making them fail. I'll work on that later. If someone could confirm they still pass. 2) There are a couple duplicate lines and I've unfortunately not spent enough time reviewing how best to refine the duplication. 3) Updated code to use updated user session changes. 4) added privatekey support for tags. I wanted to go ahead and commit my changes because it'll be another day or two to review what additional unit tests need to be added/changed based on the above changes. Headed off to take a certification test tomorrow which has taken up lots of my time. Fun times. :) One thing outstanding that you mentioned is the desire to alter the ajax script to use JSON. I fully agree with you and the change needs to be made. However, because it is consistent with the other ajax calls, I hoping you'll allow me to make those changes after the 0.98.0 release. Cheers, Mark On Sat, May 14, 2011 at 7:21 PM, Mark Pemberton <mpe...@gm...> wrote: > Nope, you are 100% correct. I did edit the create-testbookmarks from 10 to > 10k when testing something and failed to return it to 10. The privatekey2 > has it back to 10. So rant accepted. :) > > On Sat, May 14, 2011 at 4:06 AM, Christian Weiske <cw...@cw...> > wrote: >> >> Hello Mark, >> >> >> > If this is the same list you are referring to, I thought we went >> > through all of them. >> I *thought* I mentioned that the create-testbookmarks script had a >> limit of 10k instead of 10 bookmarks in it, and that's what bothered >> me. But I looked through the mails and didn't find it. Consider my >> rant baseless. >> >> >> -- >> Regards/Mit freundlichen Grüßen >> Christian Weiske >> >> -=≡ Geeking around in the name of science since 1982 ≡=- >> >> >> ------------------------------------------------------------------------------ >> Achieve unprecedented app performance and reliability >> What every C/C++ and Fortran developer should know. >> Learn how Intel has extended the reach of its next-generation tools >> to help boost performance applications - inlcuding clusters. >> http://p.sf.net/sfu/intel-dev2devmay >> _______________________________________________ >> Semanticscuttle-devel mailing list >> Sem...@li... >> https://lists.sourceforge.net/lists/listinfo/semanticscuttle-devel >> > > |
|
From: Christian W. <cw...@cw...> - 2011-05-17 20:28:37
Attachments:
signature.asc
|
Hello Mark, > Unit errors: 1a) One in Bookmark2TagTest::testGetPopularTagsDays > appears to exist in master, so I can't claim that. I spent some time > trying to resolve it, but was unable to. What error do you get? I don't get any errors in master. > 1b) Another unresolved issue is in > Tag2TagTest::testSearchThroughLinkedTags which is a result of a change > I did in getBookmarks(). I'll go ahead and apologize now if I'm wrong > in my understanding of how the tag2tag's work. The change I made > seems to make sense, and seemed to produce the results that I was > expecting. However, when I saw this test fail, red flags started > raising and now my change is in doubt. So either the update to > getBookmarks() is all wrong and all of it needs to be rejected, or it > is right and the test needs to be updated. Regardless if I'm right or > wrong, I'd love to use this opportunity to properly document this area > of work to help eliminate confusion in the future. I'll have a look at it. > 1c) The rest of the unit errors are in the API and it appears to be my > environment that is making them fail. I'll work on that later. If > someone could confirm they still pass. As I said, everyt single test works for me. > I wanted to go ahead and commit my changes because it'll be another > day or two to review what additional unit tests need to be > added/changed based on the above changes. Headed off to take a > certification test tomorrow which has taken up lots of my time. Fun > times. :) Good luck! > One thing outstanding that you mentioned is the desire to alter the > ajax script to use JSON. I fully agree with you and the change needs > to be made. However, because it is consistent with the other ajax > calls, I hoping you'll allow me to make those changes after the 0.98.0 > release. Fine with me. -- Regards/Mit freundlichen Grüßen Christian Weiske -=≡ Geeking around in the name of science since 1982 ≡=- |
|
From: Christian W. <cw...@cw...> - 2011-05-18 06:22:29
Attachments:
signature.asc
|
Hello Mark, > in my understanding of how the tag2tag's work. The change I made > seems to make sense, and seemed to produce the results that I was > expecting. However, when I saw this test fail, red flags started > raising and now my change is in doubt. So either the update to > getBookmarks() is all wrong and all of it needs to be rejected, or it > is right and the test needs to be updated. Regardless if I'm right or > wrong, I'd love to use this opportunity to properly document this area > of work to help eliminate confusion in the future. Your change to getBookmarks related to tags (920f7fc623ecad4f1338ab68326f2817c12c4610) does not take into account that several tags may be found. You're always checking on tag table 0 and don't add any other tag table joins. If the first tag matches, the second obviously cannot match, which is why it fails. -- Regards/Mit freundlichen Grüßen Christian Weiske -=≡ Geeking around in the name of science since 1982 ≡=- |
|
From: Christian W. <cw...@cw...> - 2011-05-18 06:31:17
Attachments:
signature.asc
|
Hello again, > > in my understanding of how the tag2tag's work. The change I made > > seems to make sense, and seemed to produce the results that I was > > expecting. However, when I saw this test fail, red flags started > > raising and now my change is in doubt. So either the update to > > getBookmarks() is all wrong and all of it needs to be rejected, or > > it is right and the test needs to be updated. Regardless if I'm > > right or wrong, I'd love to use this opportunity to properly > > document this area of work to help eliminate confusion in the > > future. > Your change to getBookmarks related to tags > (920f7fc623ecad4f1338ab68326f2817c12c4610) does not take into account > that several tags may be found. You're always checking on tag table 0 > and don't add any other tag table joins. If the first tag matches, the > second obviously cannot match, which is why it fails. It'd be cool if you'd change the privatekey things in the privatekey branch only. Branches are cheap and make it easier for me to review and merge the relevant changes only. If it's broken it must be fixed of course, but in this case things were not broken as far as I know. Just open a new branch for such changes. -- Regards/Mit freundlichen Grüßen Christian Weiske -=≡ Geeking around in the name of science since 1982 ≡=- |
|
From: Mark P. <mpe...@gm...> - 2011-05-18 13:54:24
|
Just figured why we're not on the same page. I'm thinking the use of multiple tags is an inclusive search (i.e. give me bookmarks that have a tag "cat" or "dog") whereas you are wanting it to be an exclusive search (i.e. give me bookmarks that have both "cat" and "dog" as tags). I'll change getBookmarks() back to the way it was. On a side note, I don't see the need in this case to have multiple joins to the same table when multiple tags are searched. But since the original code seems to work well, I'm not going to change it. :) On Wed, May 18, 2011 at 1:58 AM, Christian Weiske <cw...@cw...> wrote: > Hello Mark, > > > >> in my understanding of how the tag2tag's work. The change I made >> seems to make sense, and seemed to produce the results that I was >> expecting. However, when I saw this test fail, red flags started >> raising and now my change is in doubt. So either the update to >> getBookmarks() is all wrong and all of it needs to be rejected, or it >> is right and the test needs to be updated. Regardless if I'm right or >> wrong, I'd love to use this opportunity to properly document this area >> of work to help eliminate confusion in the future. > > Your change to getBookmarks related to tags > (920f7fc623ecad4f1338ab68326f2817c12c4610) does not take into account > that several tags may be found. You're always checking on tag table 0 > and don't add any other tag table joins. If the first tag matches, the > second obviously cannot match, which is why it fails. > > > -- > Regards/Mit freundlichen Grüßen > Christian Weiske > > -=≡ Geeking around in the name of science since 1982 ≡=- > > ------------------------------------------------------------------------------ > What Every C/C++ and Fortran developer Should Know! > Read this article and learn how Intel has extended the reach of its > next-generation tools to help Windows* and Linux* C/C++ and Fortran > developers boost performance applications - including clusters. > http://p.sf.net/sfu/intel-dev2devmay > _______________________________________________ > Semanticscuttle-devel mailing list > Sem...@li... > https://lists.sourceforge.net/lists/listinfo/semanticscuttle-devel > > |
|
From: Christian W. <cw...@cw...> - 2011-05-19 05:32:45
Attachments:
signature.asc
|
Hello Mark, > Just figured why we're not on the same page. I'm thinking the use of > multiple tags is an inclusive search (i.e. give me bookmarks that have > a tag "cat" or "dog") whereas you are wanting it to be an exclusive > search (i.e. give me bookmarks that have both "cat" and "dog" as > tags). Yep, that's really so. The docblock also does not state what the tags parameter does, so both of us are correct. I don't like the way this method works anyway and would rather have the possibility to use complex searches (foo1 foo2 OR bar -tag:baz -user:blubb). This may be implemented by adding a new bookmarks search method, or using i.e. lucene/solr to the mix - so it's a bit off. -- Regards/Mit freundlichen Grüßen Christian Weiske -=≡ Geeking around in the name of science since 1982 ≡=- |
|
From: Mark P. <mpe...@gm...> - 2011-05-20 18:16:00
|
Ok, the getBookmarks() function is back the way it was and all tests pass. Give it a go and see if I'm missing anything. On Thu, May 19, 2011 at 12:44 AM, Christian Weiske <cw...@cw...> wrote: > Hello Mark, > > >> Just figured why we're not on the same page. I'm thinking the use of >> multiple tags is an inclusive search (i.e. give me bookmarks that have >> a tag "cat" or "dog") whereas you are wanting it to be an exclusive >> search (i.e. give me bookmarks that have both "cat" and "dog" as >> tags). > Yep, that's really so. The docblock also does not state what the tags > parameter does, so both of us are correct. I don't like the way this > method works anyway and would rather have the possibility to use > complex searches (foo1 foo2 OR bar -tag:baz -user:blubb). This may be > implemented by adding a new bookmarks search method, or using i.e. > lucene/solr to the mix - so it's a bit off. > > > -- > Regards/Mit freundlichen Grüßen > Christian Weiske > > -=≡ Geeking around in the name of science since 1982 ≡=- > > ------------------------------------------------------------------------------ > What Every C/C++ and Fortran developer Should Know! > Read this article and learn how Intel has extended the reach of its > next-generation tools to help Windows* and Linux* C/C++ and Fortran > developers boost performance applications - including clusters. > http://p.sf.net/sfu/intel-dev2devmay > _______________________________________________ > Semanticscuttle-devel mailing list > Sem...@li... > https://lists.sourceforge.net/lists/listinfo/semanticscuttle-devel > > |
|
From: Christian W. <cw...@cw...> - 2011-05-20 20:24:37
Attachments:
signature.asc
|
Hello Mark, > Ok, the getBookmarks() function is back the way it was and all tests > pass. Give it a go and see if I'm missing anything. That's the output I get: cweiske:~/Dev/semanticscuttle/cwdev/tests> phpunit . PHPUnit 3.5.10 by Sebastian Bergmann. ........F...................................................... 63 / 212 ( 29%) ...E...........S............................................... 126 / 212 ( 59%) ............................................................... 189 / 212 ( 89%) ...................S... Time: 7 seconds, Memory: 20.75Mb There was 1 error: 1) Api_OpenSearchTest::testOpenSearchAvailable simplexml_load_string(): Entity: line 10: parser error : attributes construct error /home/cweiske/Dev/semanticscuttle/cwdev/tests/Api/OpenSearchTest.php:14 -- There was 1 failure: 1) Bookmark2TagTest::testGetPopularTagsDays Failed asserting that <integer:0> matches expected <integer:3>. /home/cweiske/Dev/semanticscuttle/cwdev/tests/Bookmark2TagTest.php:292 FAILURES! Tests: 212, Assertions: 938, Failures: 1, Errors: 1, Skipped: 2. -- Regards/Mit freundlichen Grüßen Christian Weiske -=≡ Geeking around in the name of science since 1982 ≡=- |
|
From: Mark P. <mpe...@gm...> - 2011-05-21 18:47:50
|
Ok, I think I am finally in the game. All tests pass. I did have some config issues which didn't allow me to run the API tests. Want to take another look? On Fri, May 20, 2011 at 4:24 PM, Christian Weiske <cw...@cw...> wrote: > Hello Mark, > > >> Ok, the getBookmarks() function is back the way it was and all tests >> pass. Give it a go and see if I'm missing anything. > > That's the output I get: > > cweiske:~/Dev/semanticscuttle/cwdev/tests> phpunit . > PHPUnit 3.5.10 by Sebastian Bergmann. > > ........F...................................................... 63 / > 212 > ( 29%) ...E...........S............................................... > 126 / 212 > ( 59%) ............................................................... > 189 / 212 ( 89%) ...................S... > > Time: 7 seconds, Memory: 20.75Mb > > There was 1 error: > > 1) Api_OpenSearchTest::testOpenSearchAvailable > simplexml_load_string(): Entity: line 10: parser error : attributes > construct error > > /home/cweiske/Dev/semanticscuttle/cwdev/tests/Api/OpenSearchTest.php:14 > > -- > > > There was 1 failure: > > 1) Bookmark2TagTest::testGetPopularTagsDays > Failed asserting that <integer:0> matches expected <integer:3>. > > /home/cweiske/Dev/semanticscuttle/cwdev/tests/Bookmark2TagTest.php:292 > > FAILURES! > Tests: 212, Assertions: 938, Failures: 1, Errors: 1, Skipped: 2. > > -- > Regards/Mit freundlichen Grüßen > Christian Weiske > > -=≡ Geeking around in the name of science since 1982 ≡=- > > ------------------------------------------------------------------------------ > What Every C/C++ and Fortran developer Should Know! > Read this article and learn how Intel has extended the reach of its > next-generation tools to help Windows* and Linux* C/C++ and Fortran > developers boost performance applications - including clusters. > http://p.sf.net/sfu/intel-dev2devmay > _______________________________________________ > Semanticscuttle-devel mailing list > Sem...@li... > https://lists.sourceforge.net/lists/listinfo/semanticscuttle-devel > > |
|
From: Christian W. <cw...@cw...> - 2011-05-21 22:30:59
Attachments:
signature.asc
|
Hello Mark,
> Ok, I think I am finally in the game. All tests pass. I did have
> some config issues which didn't allow me to run the API tests. Want
> to take another look?
Strangely, I do get one single failed test that I don't get in master.
Btw, please merge master into your branch.
I'll review it today or tomorrow nontheless.
cweiske:~/Dev/semanticscuttle/cwdev/tests> phpunit .
PHPUnit 3.5.10 by Sebastian Bergmann.
........F...................................................... 63 /
212
( 29%) ...............S...............................................
126 / 212
( 59%) ...............................................................
189 / 212 ( 89%) ...................S...
Time: 7 seconds, Memory: 20.50Mb
There was 1 failure:
1) Bookmark2TagTest::testGetPopularTagsDays
Failed asserting that an array contains
Array
(
[tag] => one
[bCount] => 3
)
.
/home/cweiske/Dev/semanticscuttle/cwdev/tests/Bookmark2TagTest.php:293
FAILURES!
Tests: 212, Assertions: 944, Failures: 1, Skipped: 2.
--
Regards/Mit freundlichen Grüßen
Christian Weiske
-=≡ Geeking around in the name of science since 1982 ≡=-
|
|
From: Mark P. <mpe...@gm...> - 2011-05-22 02:14:56
|
I knew this was going to be an issue because I remember it was working for you earlier, but not for me. I changed it and it worked for me, but stopped working for you. Looked at it again and made some final tweaks. It ended up being a two part issue - partly a timezone issue, and partly a strtodate() issue. When "today" is called, it doesn't specify time. So I changed it to "now". Additionally, the 24 hours ago plus 1 hour (which is to equal 25 hours ago) was incorrect. Needed a -1 on the hour as well. Anyway, I think we're good now. Give it a look tomorrow. And, merged with master too. :) Cheers, Mark On Sat, May 21, 2011 at 6:30 PM, Christian Weiske <cw...@cw...> wrote: > Hello Mark, > > >> Ok, I think I am finally in the game. All tests pass. I did have >> some config issues which didn't allow me to run the API tests. Want >> to take another look? > > Strangely, I do get one single failed test that I don't get in master. > Btw, please merge master into your branch. > > I'll review it today or tomorrow nontheless. > > cweiske:~/Dev/semanticscuttle/cwdev/tests> phpunit . > PHPUnit 3.5.10 by Sebastian Bergmann. > > ........F...................................................... 63 / > 212 > ( 29%) ...............S............................................... > 126 / 212 > ( 59%) ............................................................... > 189 / 212 ( 89%) ...................S... > > Time: 7 seconds, Memory: 20.50Mb > > There was 1 failure: > > 1) Bookmark2TagTest::testGetPopularTagsDays > Failed asserting that an array contains > Array > ( > [tag] => one > [bCount] => 3 > ) > . > > /home/cweiske/Dev/semanticscuttle/cwdev/tests/Bookmark2TagTest.php:293 > > FAILURES! > Tests: 212, Assertions: 944, Failures: 1, Skipped: 2. > > > -- > Regards/Mit freundlichen Grüßen > Christian Weiske > > -=≡ Geeking around in the name of science since 1982 ≡=- > > ------------------------------------------------------------------------------ > What Every C/C++ and Fortran developer Should Know! > Read this article and learn how Intel has extended the reach of its > next-generation tools to help Windows* and Linux* C/C++ and Fortran > developers boost performance applications - including clusters. > http://p.sf.net/sfu/intel-dev2devmay > _______________________________________________ > Semanticscuttle-devel mailing list > Sem...@li... > https://lists.sourceforge.net/lists/listinfo/semanticscuttle-devel > > |
|
From: Christian W. <cw...@cw...> - 2011-05-22 08:41:27
Attachments:
signature.asc
|
Hello Mark, > I knew this was going to be an issue because I remember it was working > for you earlier, but not for me. I changed it and it worked for me, > but stopped working for you. Looked at it again and made some final > tweaks. It ended up being a two part issue - partly a timezone issue, > and partly a strtodate() issue. When "today" is called, it doesn't > specify time. So I changed it to "now". Additionally, the 24 hours > ago plus 1 hour (which is to equal 25 hours ago) was incorrect. > Needed a -1 on the hour as well. Anyway, I think we're good now. > Give it a look tomorrow. 1. The tests run now for me, too - but I'm not happy with the change you introduced, and that is changing from "gmdate" to "date". Using gmdate returns the UTC time, which is awesome because we can move to another server in a different location on the world and still have the same outputs. Also, with UTC dates, we can localize the dates for the current time zone of the user without problems. I saw that you changed gmdate to date in src/SemanticScuttle/Service/Bookmark.php, which is something that should not be necessary. The rest of the code still has gmdate anyway. 2. In www/ajaxGetNewPrivateKey.php, please use the new header as you see it from i.e. www/api/posts_add.php, and please use AGPL v3 or later as license if you're ok with that. The AGPL (Affero GPL) plugs a hole in the GPL license that occurs when you're distributing code for web apps. The GPL requires you to contribute the changes back to the public when you distribute the software to the customer. With web apps, this mostly does never happen since the software keeps to be installed on the server only. With AGPL, people need to contribute their changes to the code back when they make the software available, that is if they publicize the web app installation - which is better suited for us. 3. There is no docblock for getNewPrivateKey in www/jsScuttle.php 4. www/rss.php still has $isTempLogin, which we don't want (see the mail I explained about crashes and so). loginPrivateKey in src/SemanticScuttle/Service/User.php writes into the session instead of using the revamped setCurrentUserId(). 5. I'd like to see tests that verify the private rss links are available in the header everywhere they should occur. You should do that in tests/www/$relevantfile.php, loading the page with SimpleXML and doing an xpath query on it to find the rss feed. Examples for that are in Api/OpenSearchTest.php and www/bookmarksTest.php. Your query should be like 6. The title of the feed on the main page is "$page: (private) $username" instead of "Recent bookmarks (+private)". IIRC the feed there should show all bookmarks + the user's private/shared ones. -- Regards/Mit freundlichen Grüßen Christian Weiske -=≡ Geeking around in the name of science since 1982 ≡=- |
|
From: Christian W. <cw...@cw...> - 2011-05-29 19:56:58
Attachments:
signature.asc
|
Hello Mark, Any update yet? I didn't see a commit and I really want to get 0.98 out. -- Regards/Mit freundlichen Grüßen Christian Weiske -=≡ Geeking around in the name of science since 1982 ≡=- |
|
From: Mark P. <mpe...@gm...> - 2011-05-30 03:21:26
|
Sorry, went on vacation with intentions to finish up, but server at home locked up (thanks to failing hardware). I'm home now and have tomorrow (Monday) off and should finish up with the last few things you've mentioned. Thanks for being patient. On Sun, May 29, 2011 at 3:56 PM, Christian Weiske <cw...@cw...> wrote: > Hello Mark, > > > Any update yet? I didn't see a commit and I really want to get 0.98 out. > > -- > Regards/Mit freundlichen Grüßen > Christian Weiske > > -=≡ Geeking around in the name of science since 1982 ≡=- > > ------------------------------------------------------------------------------ > vRanger cuts backup time in half-while increasing security. > With the market-leading solution for virtual backup and recovery, > you get blazing-fast, flexible, and affordable data protection. > Download your free trial now. > http://p.sf.net/sfu/quest-d2dcopy1 > _______________________________________________ > Semanticscuttle-devel mailing list > Sem...@li... > https://lists.sourceforge.net/lists/listinfo/semanticscuttle-devel > > |
|
From: Christian W. <cw...@cw...> - 2011-05-30 04:12:44
Attachments:
signature.asc
|
Hello Mark, > Sorry, went on vacation with intentions to finish up, but server at > home locked up (thanks to failing hardware). I'm home now and have > tomorrow (Monday) off and should finish up with the last few things > you've mentioned. Thanks for being patient. If it's too much stress for you, we can target privatekey for 0.99, but we can also wait. -- Regards/Mit freundlichen Grüßen Christian Weiske -=≡ Geeking around in the name of science since 1982 ≡=- |
|
From: Mark P. <mpe...@gm...> - 2011-05-31 03:40:25
|
1. Understood. I'll revert it back to gmdate. I'll also try to find where the breakdown is on my side why it didn't work for me originally. 2. Understood. 3. Added. 4. Updated and testing my changes now. 5. Working on tests now. 6. Is "Recent Bookmarks (+private)" adequate? Is there any reason for the user name to be on there? On Sun, May 22, 2011 at 4:41 AM, Christian Weiske <cw...@cw...> wrote: > Hello Mark, > > >> I knew this was going to be an issue because I remember it was working >> for you earlier, but not for me. I changed it and it worked for me, >> but stopped working for you. Looked at it again and made some final >> tweaks. It ended up being a two part issue - partly a timezone issue, >> and partly a strtodate() issue. When "today" is called, it doesn't >> specify time. So I changed it to "now". Additionally, the 24 hours >> ago plus 1 hour (which is to equal 25 hours ago) was incorrect. >> Needed a -1 on the hour as well. Anyway, I think we're good now. >> Give it a look tomorrow. > > 1. The tests run now for me, too - but I'm not happy with the change you > introduced, and that is changing from "gmdate" to "date". Using gmdate > returns the UTC time, which is awesome because we can move to > another server in a different location on the world and still have the > same outputs. Also, with UTC dates, we can localize the dates for the > current time zone of the user without problems. > I saw that you changed gmdate to > date in src/SemanticScuttle/Service/Bookmark.php, which is something > that should not be necessary. The rest of the code still has gmdate > anyway. > > 2. In www/ajaxGetNewPrivateKey.php, please use the new > header as you see it from i.e. www/api/posts_add.php, and please use > AGPL v3 or later as license if you're ok with that. The AGPL (Affero > GPL) plugs a hole in the GPL license that occurs when you're > distributing code for web apps. The GPL requires you to contribute the > changes back to the public when you distribute the software to the > customer. With web apps, this mostly does never happen since the > software keeps to be installed on the server only. With AGPL, people > need to contribute their changes to the code back when they make the > software available, that is if they publicize the web app installation > - which is better suited for us. > > 3. There is no docblock for getNewPrivateKey in www/jsScuttle.php > > 4. www/rss.php still has $isTempLogin, which we don't want (see the > mail I explained about crashes and so). loginPrivateKey in > src/SemanticScuttle/Service/User.php writes into the session instead of > using the revamped setCurrentUserId(). > > 5. I'd like to see tests that verify the private rss links are > available in the header everywhere they should occur. You should do > that in tests/www/$relevantfile.php, loading the page with SimpleXML > and doing an xpath query on it to find the rss feed. Examples for that > are in Api/OpenSearchTest.php and www/bookmarksTest.php. Your query > should be like > > 6. The title of the feed on the main page is "$page: (private) > $username" instead of "Recent bookmarks (+private)". IIRC the feed > there should show all bookmarks + the user's private/shared ones. > > -- > Regards/Mit freundlichen Grüßen > Christian Weiske > > -=≡ Geeking around in the name of science since 1982 ≡=- > > ------------------------------------------------------------------------------ > What Every C/C++ and Fortran developer Should Know! > Read this article and learn how Intel has extended the reach of its > next-generation tools to help Windows* and Linux* C/C++ and Fortran > developers boost performance applications - including clusters. > http://p.sf.net/sfu/intel-dev2devmay > _______________________________________________ > Semanticscuttle-devel mailing list > Sem...@li... > https://lists.sourceforge.net/lists/listinfo/semanticscuttle-devel > > |
|
From: Christian W. <cw...@cw...> - 2011-06-07 19:10:29
Attachments:
signature.asc
|
Hello Mark, > 1. Understood. I'll revert it back to gmdate. I'll also try to find > where the breakdown is on my side why it didn't work for me > originally. So what was it? > 6. Is "Recent Bookmarks (+private)" adequate? Is there any reason for > the user name to be on there? Yes, because you might subscribe to the official bookmarks with your private key and your own private bookmarks. A different title helps here. -- Regards/Mit freundlichen Grüßen Christian Weiske -=≡ Geeking around in the name of science since 1982 ≡=- |
|
From: Mark P. <mpe...@gm...> - 2011-06-04 04:44:51
|
Didn't get a reply from my previous e-mail, so I went ahead and made the assumption to #6 and everything else should be ok. Forgive me on my poor response, just been crazy after vacation. Also, historically I've not done much unit testing, but as I'm learning through this the unit testing can prove quite valuable. I think I've added what you requested, so let me know if you see anything else. I'm a bit of a novice on xpath but was able put it together. I've also merged with master. :) Cheers all! Mark On Mon, May 30, 2011 at 11:40 PM, Mark Pemberton <mpe...@gm...> wrote: > 1. Understood. I'll revert it back to gmdate. I'll also try to find > where the breakdown is on my side why it didn't work for me > originally. > 2. Understood. > 3. Added. > 4. Updated and testing my changes now. > 5. Working on tests now. > 6. Is "Recent Bookmarks (+private)" adequate? Is there any reason for > the user name to be on there? > > > On Sun, May 22, 2011 at 4:41 AM, Christian Weiske <cw...@cw...> wrote: >> Hello Mark, >> >> >>> I knew this was going to be an issue because I remember it was working >>> for you earlier, but not for me. I changed it and it worked for me, >>> but stopped working for you. Looked at it again and made some final >>> tweaks. It ended up being a two part issue - partly a timezone issue, >>> and partly a strtodate() issue. When "today" is called, it doesn't >>> specify time. So I changed it to "now". Additionally, the 24 hours >>> ago plus 1 hour (which is to equal 25 hours ago) was incorrect. >>> Needed a -1 on the hour as well. Anyway, I think we're good now. >>> Give it a look tomorrow. >> >> 1. The tests run now for me, too - but I'm not happy with the change you >> introduced, and that is changing from "gmdate" to "date". Using gmdate >> returns the UTC time, which is awesome because we can move to >> another server in a different location on the world and still have the >> same outputs. Also, with UTC dates, we can localize the dates for the >> current time zone of the user without problems. >> I saw that you changed gmdate to >> date in src/SemanticScuttle/Service/Bookmark.php, which is something >> that should not be necessary. The rest of the code still has gmdate >> anyway. >> >> 2. In www/ajaxGetNewPrivateKey.php, please use the new >> header as you see it from i.e. www/api/posts_add.php, and please use >> AGPL v3 or later as license if you're ok with that. The AGPL (Affero >> GPL) plugs a hole in the GPL license that occurs when you're >> distributing code for web apps. The GPL requires you to contribute the >> changes back to the public when you distribute the software to the >> customer. With web apps, this mostly does never happen since the >> software keeps to be installed on the server only. With AGPL, people >> need to contribute their changes to the code back when they make the >> software available, that is if they publicize the web app installation >> - which is better suited for us. >> >> 3. There is no docblock for getNewPrivateKey in www/jsScuttle.php >> >> 4. www/rss.php still has $isTempLogin, which we don't want (see the >> mail I explained about crashes and so). loginPrivateKey in >> src/SemanticScuttle/Service/User.php writes into the session instead of >> using the revamped setCurrentUserId(). >> >> 5. I'd like to see tests that verify the private rss links are >> available in the header everywhere they should occur. You should do >> that in tests/www/$relevantfile.php, loading the page with SimpleXML >> and doing an xpath query on it to find the rss feed. Examples for that >> are in Api/OpenSearchTest.php and www/bookmarksTest.php. Your query >> should be like >> >> 6. The title of the feed on the main page is "$page: (private) >> $username" instead of "Recent bookmarks (+private)". IIRC the feed >> there should show all bookmarks + the user's private/shared ones. >> >> -- >> Regards/Mit freundlichen Grüßen >> Christian Weiske >> >> -=≡ Geeking around in the name of science since 1982 ≡=- >> >> ------------------------------------------------------------------------------ >> What Every C/C++ and Fortran developer Should Know! >> Read this article and learn how Intel has extended the reach of its >> next-generation tools to help Windows* and Linux* C/C++ and Fortran >> developers boost performance applications - including clusters. >> http://p.sf.net/sfu/intel-dev2devmay >> _______________________________________________ >> Semanticscuttle-devel mailing list >> Sem...@li... >> https://lists.sourceforge.net/lists/listinfo/semanticscuttle-devel >> >> > |
×
Want the latest updates on software, tech news, and AI?
Get latest updates about software, tech news, and AI from SourceForge directly in your inbox once a month.
Thanks for helping keep SourceForge clean.
X