You can subscribe to this list here.
2002 |
Jan
|
Feb
|
Mar
|
Apr
|
May
|
Jun
|
Jul
|
Aug
|
Sep
|
Oct
|
Nov
|
Dec
(3) |
---|---|---|---|---|---|---|---|---|---|---|---|---|
2003 |
Jan
(1) |
Feb
|
Mar
(1) |
Apr
(8) |
May
(10) |
Jun
(15) |
Jul
(32) |
Aug
(10) |
Sep
(13) |
Oct
(13) |
Nov
(1) |
Dec
|
2004 |
Jan
|
Feb
|
Mar
(1) |
Apr
(1) |
May
(3) |
Jun
|
Jul
(5) |
Aug
(9) |
Sep
(17) |
Oct
(1) |
Nov
|
Dec
(1) |
2005 |
Jan
(2) |
Feb
(17) |
Mar
|
Apr
|
May
|
Jun
|
Jul
|
Aug
|
Sep
|
Oct
(6) |
Nov
|
Dec
|
2006 |
Jan
|
Feb
|
Mar
(8) |
Apr
(1) |
May
|
Jun
(1) |
Jul
|
Aug
(4) |
Sep
|
Oct
|
Nov
|
Dec
|
From: Shane <j.s...@gm...> - 2006-08-11 10:07:49
|
Thanks Kevin for getting back to me and thanks for the info. Even though I do want to see the Python version, I can't take the time to take a look now. int On 8/10/06, Kevin Dwyer <ke...@ph...> wrote: > > On Mon, Aug 07, 2006 at 08:17:55AM -0400, Shane wrote: > > Hello. > > I am thinking about making some changes to bidwatcher and wanted to get > the > > latest info first. I have browsed the project site and some source code > and > > have not seen any significant changes since early last year. Is that > change > > that was made? Is the 1.3.17 version the latest? Is there a new > release in > > the works now? > > Like Klaus said, 1.3.17 is broken. If you're going to work on code, by > all means work off of the CVS code. > > I've been too busy to really tackle the problems with the code, and > quite frankly it's a lot of tedious work to get bidding working again. > I did make a python version of bidwatcher that can do almost everything > the C++ version can do, and it's easier to change. I can send you the > latest copy if you're interested in playing. > > -kpd > |
From: Kevin D. <ke...@ph...> - 2006-08-10 23:01:25
|
On Mon, Aug 07, 2006 at 08:17:55AM -0400, Shane wrote: > Hello. > I am thinking about making some changes to bidwatcher and wanted to get the > latest info first. I have browsed the project site and some source code and > have not seen any significant changes since early last year. Is that change > that was made? Is the 1.3.17 version the latest? Is there a new release in > the works now? Like Klaus said, 1.3.17 is broken. If you're going to work on code, by all means work off of the CVS code. I've been too busy to really tackle the problems with the code, and quite frankly it's a lot of tedious work to get bidding working again. I did make a python version of bidwatcher that can do almost everything the C++ version can do, and it's easier to change. I can send you the latest copy if you're interested in playing. -kpd |
From: Klaus S. <kl...@si...> - 2006-08-08 15:00:07
|
Shane wrote: > Hello. > I am thinking about making some changes to bidwatcher and wanted to get the > latest info first. I have browsed the project site and some source code and > have not seen any significant changes since early last year. Is that change > that was made? Is the 1.3.17 version the latest? Is there a new release in > the works now? Take the latest CVS snapshot please. 1.3.17 is outdated. I worked on the CVS in the last year: after changes of the layout at the ebay servers. > I have not fetched the latest copy of the source (for reasons I will touch > on below) so I'm not being lazy by asking if the current version is up to > date and works. Does it? I did look at the patches and bugs on SF and did > not see anything about it not working or being broken by one of eBay's > recent changes. The latest version (CVS) should work -- at least the watching and logging ability of articles. I don't use the bidding feature, so I'm not able to speak about it. > Are there any new directions in the project or new features that someone is > actively working on now? I'm not aware of active developing on new features. I'm just fixing/maintaining it, if the layout changes and the parsing needs some fixes. > I'm having a little trouble with my linux server and my connectivity is > extremely unreliable right now so while I am working to correct these > problems I do not expect to make any changes. It will probably be 1-3 weeks > before I can really make any progress. Good luck. Regards, Klaus (aka singvogel). -- Klaus Singvogel GnuPG-Key-ID: 1024R/5068792D 1994-06-27 |
From: Shane <j.s...@gm...> - 2006-08-07 12:17:56
|
Hello. I am thinking about making some changes to bidwatcher and wanted to get the latest info first. I have browsed the project site and some source code and have not seen any significant changes since early last year. Is that change that was made? Is the 1.3.17 version the latest? Is there a new release in the works now? I have not fetched the latest copy of the source (for reasons I will touch on below) so I'm not being lazy by asking if the current version is up to date and works. Does it? I did look at the patches and bugs on SF and did not see anything about it not working or being broken by one of eBay's recent changes. Are there any new directions in the project or new features that someone is actively working on now? I'm having a little trouble with my linux server and my connectivity is extremely unreliable right now so while I am working to correct these problems I do not expect to make any changes. It will probably be 1-3 weeks before I can really make any progress. int_80h |
From: Klaus S. <sou...@si...> - 2006-06-27 09:18:01
|
Hello everybody, eBay is doing some changes in the "end time" of the auctions today. Therefore half of the request had correct end times, and the other half hadn't: they are already ended for bidwatcher in any such case. Be aware that there is an issue in bidwatcher, and please note that I fixed it, and checked the changes into the CVS at sf.net. Regards, Klaus. -- Klaus Singvogel GnuPG-Key-ID: 1024R/5068792D 1994-06-27 |
From: Klaus S. <sou...@si...> - 2006-04-05 11:46:23
|
Hello everyone, I noticed a problem in the "end time" determination in bidwatcher (File: helpers.cpp). I realized this problem already last week. But the developer access to SourceForge didn't work, so I couldn't update the files till now. If anyone wants to update: the problem should be fixed now. Regards, Klaus. -- Klaus Singvogel GnuPG-Key-ID: 1024R/5068792D 1994-06-27 |
From: Klaus S. <sou...@si...> - 2006-03-23 16:54:20
|
Kevin, thanks for applying latest patches to the CVS. They seem to be complete now. But this week eBay changed again the layout, and another patch is necessary (see attachement). As you still didn't gave me write access to the CVS, can you please add it, please? Thanks in advance. Regards, Klaus. Kevin Dwyer wrote: > Klaus, >=20 > (I don't know if you're on the cvs list, but) I checked in several of > your patches as well as the most recent one. Let me know if I missed > anything. >=20 > -kpd >=20 >=20 >=20 > ------------------------------------------------------- > This SF.Net email is sponsored by xPML, a groundbreaking scripting lang= uage > that extends applications into web and mobile media. Attend the live we= bcast > and join the prime developer group breaking into this new coding territ= ory! > http://sel.as-us.falkag.net/sel?cmd=3Dlnk&kid=3D110944&bid=3D241720&dat= =3D121642 > _______________________________________________ > Bidwatcher-devel mailing list > Bid...@li... > https://lists.sourceforge.net/lists/listinfo/bidwatcher-devel --=20 AOL und Yahoo-Kunden aufgepasst! Wer zuk=FCnftig E-Mail direkt zustellen will, kann diesen Service f=FCr 2= Cent pro E-Mail erhalten. AOL- und Yahoo-Kunden, die dies nicht machen, d=FCrf= en sich nicht mehr wundern, wenn deren E-Mail ungelesen in den Spam-Folder wandern. Seri=F6s gemeinte Anfragen erhalten die Kontodaten auf Nachfrage= . Gegenreaktion auf: http://www.heise.de/newsticker/meldung/69232 Klaus Singvogel GnuPG-Key-ID: 1024R/5068792D 1994-06-27 |
From: Kevin D. <ke...@ph...> - 2006-03-14 12:43:30
|
Klaus, (I don't know if you're on the cvs list, but) I checked in several of your patches as well as the most recent one. Let me know if I missed anything. -kpd |
From: Kevin D. <ke...@ph...> - 2006-03-14 12:41:58
|
I updated my python source to handle the new changes, and I also added some code to login to eBay (which means that bidding should be possible with just a little more fiddling.) The neat thing is that this is only like 500 lines of python and it's practically done. Here it is in case anyone is interested in playing with it. -kpd |
From: Klaus S. <sou...@si...> - 2006-03-11 10:57:16
|
Kevin, I had a look at the patch and the broken part only works, if you apply another patch, which I already sent to the bidwatcher-devel mailinglist at 2005-10-13 05:28; named bidwatcher-1.3.17-account.patch You might find it in your mail archive or at the archive at sourceforge. :-) If you don't find a applyable copy, send me a private e-mail please, and I'll re-send him to you directly (I don't want to bloat the archive with the same things again). To point it out: the patch is really repairing things, and isn't a feature enhacement... If you didn't apply that patch, I asume that you didn't apply the multilple spaces fix neither? (I sent it to the ML a few minutes later) Whereas the account patch is only a minor fix, which occurs only on rare circumstances, the later is absolut necessary to get bidwatcher still running. Otherwise eBay tokens are no longer found. Please grab the bidwatcher-1.3.17-multiple_space.patch and apply it to the CVS repository as well. Thanks in advance. Regards, Klaus. Kevin Dwyer wrote: > Klaus, >=20 > Would you mind reworking your patch so that it is against the last > revision of the 1.3 branch? It almost applies, but some hunks fail. > Output below: >=20 > $ patch -p1 < bidwatcher-1.3.17-changes_20060307.patch > patching file helpers.cpp > Hunk #3 succeeded at 1295 (offset -14 lines). > Hunk #5 succeeded at 1495 (offset -12 lines). > Hunk #6 FAILED at 1608. > 1 out of 6 hunks FAILED -- saving rejects to file helpers.cpp.rej > patching file bidwatcher.cpp > Hunk #1 succeeded at 4044 with fuzz 2 (offset -3 lines). > Hunk #2 succeeded at 4374 with fuzz 2 (offset -2 lines). >=20 > Thanks, > -kpd >=20 --=20 AOL und Yahoo-Kunden aufgepasst! Wer zuk=FCnftig E-Mail direkt zustellen will, kann diesen Service f=FCr 2= Cent pro E-Mail erhalten. AOL- und Yahoo-Kunden, die dies nicht machen, d=FCrf= en sich nicht mehr wundern, wenn deren E-Mail ungelesen in den Spam-Folder wandern. Seri=F6s gemeinte Anfragen erhalten die Kontodaten auf Nachfrage= . Gegenreaktion auf: http://www.heise.de/newsticker/meldung/69232 Klaus Singvogel GnuPG-Key-ID: 1024R/5068792D 1994-06-27 |
From: Kevin D. <ke...@ph...> - 2006-03-11 02:38:24
|
Klaus, Would you mind reworking your patch so that it is against the last revision of the 1.3 branch? It almost applies, but some hunks fail. Output below: $ patch -p1 < bidwatcher-1.3.17-changes_20060307.patch patching file helpers.cpp Hunk #3 succeeded at 1295 (offset -14 lines). Hunk #5 succeeded at 1495 (offset -12 lines). Hunk #6 FAILED at 1608. 1 out of 6 hunks FAILED -- saving rejects to file helpers.cpp.rej patching file bidwatcher.cpp Hunk #1 succeeded at 4044 with fuzz 2 (offset -3 lines). Hunk #2 succeeded at 4374 with fuzz 2 (offset -2 lines). Thanks, -kpd On Wed, Mar 08, 2006 at 05:04:01PM +0100, Klaus Singvogel wrote: > Hi, > after doing my first round of patches (till late in the night :-), > I noticed today, that it wasn't complete. Some old "hacks" bit me > now. > > So I reworked on it and finally have now a version, which satisfies my > wishes. :-) I attached it. Please drop (remove) the patches from > yesterday and use this one instead, please. > > Hopefully someone with CVS write access can patch the SourceForge > repository. > > Regards, > Klaus. > -- > Klaus Singvogel > GnuPG-Key-ID: 1024R/5068792D 1994-06-27 > --- bidwatcher-1.3.17/helpers.cpp.orig 2006-03-08 15:23:32.000000000 +0100 > +++ bidwatcher-1.3.17/helpers.cpp 2006-03-08 15:23:54.000000000 +0100 > @@ -55,7 +55,7 @@ > > const char * const CheckPrice="0123456789.$,"; // Characters allowed in a price > > -const char * const SearchOps[2][37]={{ "Starts at", // 1 > +const char * const SearchOps[2][40]={{ "Starts at", // 1 > "Currently", // 2 > "First bid", // 3 > "Quantity", // 4 > @@ -85,14 +85,17 @@ > // for some reason theBuyItNow/Purchase only > // auction have a space in front of "Price". > " Price", // 28 > - "Feedback Score:", // 29 > - "(", // 30 > + "Feedback Score:", // 29 > + "(", // 30 > "Ended:", // 31 > "Shipping and handling:", // 32 > "Shipping Cost Services Available", // 33, newer version of above > - " price:", // 34 > - "price:", // 35 > - "Seller information", // 36 > + " price:", // 34 > + "price:", // 35 > + "Seller information", // 36 > + "Score:", // 37 > + "Seller:", // 38 > + "End time:", // 39 > "*"}, > /* eBay Motors */ > { "Currently", // 1 > @@ -1306,7 +1309,8 @@ > > case 3: // first bid > case 30: > - if (strstr(LineBuffer.buf(), "(PayPal account required)")) break; > + if (!strstr(LineBuffer.buf(), "(Approximately")) > + break; > if (!FirstBid && !first_flag && strnzcmp(LineBuffer.buf(), "(Approximately")) { > if (!dutch_flag) { > DPRINTF(DLOW, ("LDFB: %s\n", LineBuffer.buf())); > @@ -1437,11 +1441,27 @@ > case 9: // time ending > if (!Ends[0]) { > //streamBuff.getline(LineBuffer.buf(), LineBuffer.size(), '\n'); > - //DPRINTF(DLOW, ("LDE: %s\n", LineBuffer.buf())); > + //DPRINTF(DLOW, ("LDE: %s\n", LineBuffer.buf())); > STRNZCPY(Ends, LineBuffer.buf() + strlen("Ends ")); > - translate_date(Ends); > - EndsValue = CalcEndsValue(Ends); > - } > + translate_date(Ends); > + EndsValue = CalcEndsValue(Ends); > + } > + case 39: // "End time:" > + // has 2 possibilities of string: > + // 1. "End time:10 hours 7 mins" > + // " (Mar-08-06 12:52:47 PST)" > + // 2. "End time: Mar-08-06 12:52:47 PST" > + // "2 days 10 hours 7 mins" > + if (!Ends[0]) { > + if (strstr(LineBuffer.buf(), " PST ")) { > + STRNZCPY(Ends, LineBuffer.buf() + strlen("End time:")); > + } else { > + streamBuff.getline(LineBuffer.buf(), LineBuffer.size(), '\n'); > + STRNZCPY(Ends, LineBuffer.buf()+2); // skip starting 2 chars " (" > + } > + translate_date(Ends); > + EndsValue = CalcEndsValue(Ends); > + } > break; > case 31: // ended > if (!Ends[0]) { > @@ -1487,6 +1507,98 @@ > SellerRate[count+1] = '\0'; > } > break; > + case 38: // Seller: > + if (!Seller[0]) { > + char *p; > + p = strchr(LineBuffer.buf(), ':'); > + if (!p) > + break; > + if (*p != ':') > + break; > + p += 2; // skip ': ' > + STRNZCPY(Seller, p); > + /* strip only "(private)" away */ > + p = strstr(Seller, "(private)"); > + if (p) *p='\0'; > + > + // Bidder Rate is typicaly in next line, but without leading '(' > + streamBuff.getline(LineBuffer.buf(), LineBuffer.size(), '\n'); > + if (LineBuffer.buf()[0] == '(' && !strchr(LineBuffer.buf(), ')')) > + streamBuff.getline(LineBuffer.buf(), LineBuffer.size(), '\n'); > + if (!strchr(LineBuffer.buf(), ')')) > + break; // validity check > + SellerRate[0] = '('; > + unsigned int count; > + for (count=1; count < sizeof(SellerRate)-1; count++) > + { > + if (LineBuffer.buf()[count-1] == '\0' || LineBuffer.buf()[count-1] == ')' > + || LineBuffer.buf()[count-1] == ' ') > + break; > + SellerRate[count] = LineBuffer.buf()[count-1]; > + } > + SellerRate[count] = ')'; > + SellerRate[count+1] = '\0'; > + if (!strncmp(SellerRate, "(0)", 4)) > + STRNZCPY(Feedbackn,"NONE"); > + } > + break; > + case 37: // Score: > + // Example: Score: 1762 | 99.9% Positive > + // Example: Score: Private | 97.0% Positive > + if (!Feedbackn[0]) { > + char TmpLine[6]; > + unsigned int count=0, f_len=0; > + int limit=0; > + int i=0, j=0; > + char *p; > + > + p = strchr(LineBuffer.buf(), ':'); > + limit = strlen(p); > + if (strstr(p, "Private")) { > + strcpy(Feedbackn, "Priv. "); > + count=strlen("Private ")+1; > + f_len=strlen("Priv. ")+1; > + } else { > + while(!isdigit(*p) && *p) > + p++; > + if (*p == '\0') { > + break; > + } > + > + for (f_len=0; count<limit && f_len<sizeof(Feedbackn)-1; count++, f_len++) { > + if (p[count] == ' ') > + break; > + Feedbackn[f_len] = p[count]; > + } > + Feedbackn[f_len] = '\0'; > + } > + if (p[count+1] == '|') { > + for (i=0, j=count+3; j<limit && i<sizeof(TmpLine)-1; i++, j++) { > + if (!isdigit(p[j]) && p[j] != '.' && p[j]!='%') { > + break; > + } > + TmpLine[i] = p[j]; > + } > + } else > + DPRINTF(DLOW, ("Score: p[count+1] is not '|': >%c<, count=%d\n", p[count+1], count)); > + TmpLine[i] = '\0'; > + if (i) { > + int a1= 6; //a1 = sizeof(Feedbackn)-strlen(TmpLine)-1; > + if (f_len < a1) { > + while (f_len < a1) { /* fills spaces */ > + Feedbackn[f_len++] = ' '; > + } > + } else { > + a1 = sizeof(Feedbackn)-strlen(TmpLine)-1; > + if (f_len < a1) { // room for 1 space ? > + Feedbackn[f_len++] = ' '; > + } > + } > + strcat(Feedbackn,TmpLine); > + } > + DPRINTF(DLOW, ("Score: Feedbackn= >%s<\n", Feedbackn)); > + } > + break; > case 11: // high bidder id > case 19: > case 24: > @@ -1508,7 +1620,7 @@ > STRNZCPY(HighBidder, p); > /* Don't cut on '('! "(www.grahamcrackers.com)" is a valid ebay account */ > /* strip only "(private)" away */ > - p = strstr(Seller, "(private)"); > + p = strstr(HighBidder, "(private)"); > if (p) *p='\0'; > > // Bidder Rate is typicaly in next line, but without leading '(' > --- bidwatcher-1.3.17/bidwatcher.cpp.orig 2006-03-08 15:23:32.000000000 +0100 > +++ bidwatcher-1.3.17/bidwatcher.cpp 2006-03-08 15:23:47.000000000 +0100 > @@ -4047,6 +4047,8 @@ > theColor = &color_near; > else if ( strstr(auction[i]->Feedbackn,"NONE") != NULL) > theColor = &color_none; > + else if ( strstr(auction[i]->Feedbackn,"Priv.") != NULL) > + theColor = &color_none; > else theColor = &color_far; > // if (auction[i]->isEndless) theColor = &color_near; > > @@ -4374,6 +4376,8 @@ > theColor = &color_near; > else if ( strstr(auction[i]->Feedbackn,"NONE") != NULL) > theColor = &color_none; > + else if ( strstr(auction[i]->Feedbackn,"Priv.") != NULL) > + theColor = &color_none; > else theColor = &color_far; > // if (auction[i]->isEndless) theColor = &color_near; > |
From: Kevin D. <ke...@ph...> - 2006-03-08 16:57:40
|
Thanks Klaus. I'll try to merge these in this week. -kpd P.S.: If there is anyone else out there with patches, please submit them. I've been horribly busy with real life so I haven't had the time to develop but I can certainly spare some time to manage. On Wed, Mar 08, 2006 at 05:04:01PM +0100, Klaus Singvogel wrote: > Hi, > after doing my first round of patches (till late in the night :-), > I noticed today, that it wasn't complete. Some old "hacks" bit me > now. > > So I reworked on it and finally have now a version, which satisfies my > wishes. :-) I attached it. Please drop (remove) the patches from > yesterday and use this one instead, please. > > Hopefully someone with CVS write access can patch the SourceForge > repository. |
From: Klaus S. <sou...@si...> - 2006-03-08 16:04:21
|
Hi, after doing my first round of patches (till late in the night :-), I noticed today, that it wasn't complete. Some old "hacks" bit me now. So I reworked on it and finally have now a version, which satisfies my wishes. :-) I attached it. Please drop (remove) the patches from yesterday and use this one instead, please. Hopefully someone with CVS write access can patch the SourceForge repository. Regards, Klaus. -- Klaus Singvogel GnuPG-Key-ID: 1024R/5068792D 1994-06-27 |
From: Klaus S. <sou...@si...> - 2006-03-07 20:31:28
|
Hi, I don't know, if anybody is still active on bidwatcher? But as it worked for quiet some time (except some timeouts at ebay :), I hope there are still others outside there. :-) Today ebay changed after a long time the layout (and the keywords) of their pages. The result was that bidwatcher couldn't any longer find out the ending time, the seller, and the seller's rating. I attache a patch, which fixes that. Regards, Klaus. -- Klaus Singvogel GnuPG-Key-ID: 1024R/5068792D 1994-06-27 |
From: Klaus S. <sou...@si...> - 2005-10-13 12:39:39
|
Hello, tonight eBay did a tricky change in the layout: they added multiple spaces in the layout, which got ignored by most browsers. The bidwatcher parser stumbled on that, because the keywords didn't begin any longer at the first character of line. I experiemented a lot on this, but after a break a came up with the attached trivial patch in the read-in mechanism. :-) It's for bidwatcher-1.3.17 and maybe the line offset might not be correct for the CVS version. Regards, Klaus. -- Klaus Singvogel GnuPG-Key-ID: 1024R/5068792D 1994-06-27 |
From: Klaus S. <sou...@si...> - 2005-10-13 12:28:42
|
Hi, I noticed a problem in bidwatcher regarding the seller / winner names. It seems that "(www.grahamcrackers.com)" is a valid ebay name -- from the old days :-) -- but is ignored by bidwatcher. Attached is a patch for bidwatcher-1.3.17 to fix this. Maybe there is small offset in the lines, because I adapted bidwatcher to fullfill my personal needs, like having more auctions. So, please, don't drop it in any such case. :-) Thanks in advance. Regards, Klaus. -- Klaus Singvogel GnuPG-Key-ID: 1024R/5068792D 1994-06-27 |
From: Kevin D. <ke...@ph...> - 2005-10-13 03:31:37
|
Hello fellow hackers, I had reason to try to learn some pyGTK over the past weekend so I thought I'd give it a whirl and mock up a python version of bidwatcher. I used glade to draw a GUI and then wrote some simple parsing routines in python. The code is much simpler (well, it doesn't do everything the original bidwatcher did, yet). Anyway, I'm attaching what I did in case anyone is interested. I'll probably keep hacking on it and maybe turn it into a better replacement to the C++ monstrosity. Enjoy, -kpd |
From: Klaus S. <sou...@si...> - 2005-10-10 21:45:18
|
Hello, I updated to latest STABLE branch of bidwatcher-1.3.17 and noticed that the description isn't shown correctly. It seems that my patch from August wasn't complete applied. Sorry. Attached is a small fix to repairt this. Regards, Klaus. -- Klaus Singvogel GnuPG-Key-ID: 1024R/5068792D 1994-06-27 |
From: Kevin D. <ke...@ph...> - 2005-10-05 01:26:53
|
Another patch from Klaus. -kpd |
From: Kevin D. <ke...@ph...> - 2005-10-04 14:30:52
|
FYI, patch from Klaus. -kpd |
From: Moritz B. <bar...@gm...> - 2005-02-24 16:28:07
|
Hi, here's another small patch. In the #ifdef'd DEBUG_NETWORK code, the proxy->hostname is never printed properly. This is because the C *printf() is used on a C++ string. It was done correctly in another section of the code. I fixed it here. Please review and include if considered appropriate. Thanks! Regards, Moritz |
From: Moritz B. <bar...@gm...> - 2005-02-24 16:24:30
|
Hi, here's another patch. Though the original code is fully correct, I believe the same #define should be used for evaluating the return value of fetchURL() as is used for creating it. Please review and include if considered appropriate. Thanks! Regards, Moritz |
From: Moritz B. <bar...@gm...> - 2005-02-24 16:21:05
|
Hi, here's the first of the simple patches I promised. This is just to correct and clarify the description of the function ProcessBidSubmission(). It is currently not only incomplete, but also wrong. I ran across this while trying to interpret the return value of this function. Please review and include if considered appropriate. Thanks! Regards, Moritz |
From: Moritz B. <bar...@gm...> - 2005-02-24 16:04:51
|
Hi again, let me please correct something, before you go away debugging stuff which isn't valid. I wrote: > 5.) eBay time fetching does not succeed. > > I think this is only behind a proxy. I recently re-checked that it > works with the old code (1.3.16-jan1 or something like that). Part of this problem was my use of an old libcurl, which apparently fails to redirect correctly. This was mentioned on this list already. Now, using the newest libcurl, the redirects work. Nevertheless, eBay time is still not fetched correctly! That may have also been the cause for this: > 3.) Snipes fail. > > The last few days, I have gotten a few of these: > [2005-02-22 17:31:33] Pre-bid on XXXYYY FAILED: Unable to determine > error. This indicates a bug. I need to wait and check. Well, actually, bidwatcher now regularly segfaults with the new libcurl. Oh my gosh... :-/ Actually, I get the GTK+ asserations as attached. (The segfault is also backtraced to the line 324 gtk_label_set_text(GTK_LABEL(statuslabel), t); in bidwatcher.cpp) I wonder if this really has to do with GTK+? Moritz |
From: Moritz B. <bar...@gm...> - 2005-02-24 12:31:56
|
Hi Jan, hi list, Jan writes: > [several problems] After having written (to bidwatcher-users, I believe) a few days ago that this was the best bidwatcher for quite some time, I now see it totally differently. I'd like to list all my observations. But before I start: There soon needs to be a new release. So if you poll the users on bidwatcher-users, and it turns out that CVS works for most, I guess it should be released as a starting point, with a disclaimer about stability. Here are all my current issues: 1.) Occasional segmentation faults when sniping, sometimes also when updating items. This has been seen by many people. I have had gdb on it, but didn't know what to look for. This is what Jan is working on in the e-mail I'm replying to. 2.) Sniping sometimes forgotten. bidwatcher sometimes just rests in one of the wait intervals, and doesn't snipe. A snipe would have been successful! I need to add debug code to see in which loop it is in. This was okay very recently (e.g. since Jan updated CVS Tuesday) but that may be coincidence. I have only seen this since the libcurl updates, but bidwatcher didn't work for quite some time before. 3.) Snipes fail. The last few days, I have gotten a few of these: [2005-02-22 17:31:33] Pre-bid on XXXYYY FAILED: Unable to determine error. This indicates a bug. Actually, this indicates that the parser in ProcessBidSubmission() didn't find any of the expected strings. eBay must have updated their formats some time ago. These need to be updated. What's more: I can't even check what bidwatcher got, because it isn't storing the response in ~/.bidwatcher/error-XXXYYY.html anymore. :-( Which brings me to the next bug: 4.) No debug files created anymore. I am not seeing any snipe2_05.txt (or similar) anymore. It had still worked back on Feb 04 2005 (when a snipe was also successfull - is that the point?). I am not seeing any error-XXXYYY.html anymore (for ages now!). This might be related to the libcurl code? I need to check. These files are very important e.g. for fixing the article parsing or the response parsing in ProcessBidSubmission() (see above). 5.) eBay time fetching does not succeed. I think this is only behind a proxy. I recently re-checked that it works with the old code (1.3.16-jan1 or something like that). Luckily, my system time (and eBay's) is quite okay. 6.) Spurious characters displayed in the "Description" column. As I wrote recently, I'm often seeing extra '"'s at the end of the description. (They're not in the HTML source of the auction, I swear!) Sometimes, there's more, such as '0"'. Updating the particular auction sometimes fixes this. But I think the updates also sometimes ruin it. This might be more memory corruption? 7.) I think the intervals between updates sometimes don't correspond to what I have configured. I have seen auctions ending soon (e.g in less than 15 minutes), but bidwatcher just sits around and idles, instead of updating every 60 seconds. I need to add debug code for this as well. This is annoying, but not really harmful(?). But this may by leading to the "forgotten snipes" described above. I think I have been seeing this for a very long time. 8.) No final update of an auction. bidwatcher fails to update the auction (or the displayed values) a final time after the auction has ended. This is regardless of whether a snipe is set or not. It is very very annoying. I have definitly been seeing this for ages! Okay, that shall be enough for the moment. I will try to add some debug code, and post the debug patches and results here (and fixes if I should find them). But I'm not prepared to tackle these problems on my own, sorry. (All of us here don't have time, you see. ;->) I will also post some small patch snippets for code improvement. I would very much appreciate the CVS members (Jan and Kevin?) to review and approve them. Thanks guys, have a nice day, Moritz |