scriptonite-dev Mailing List for Scriptonite (Page 2)
Status: Planning
Brought to you by:
egagnon
Menu
▾
▴
scriptonite-dev
You can subscribe to this list here.
| 2000 |
Jan
|
Feb
(13) |
Mar
(29) |
Apr
(13) |
May
|
Jun
|
Jul
|
Aug
|
Sep
|
Oct
|
Nov
|
Dec
|
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 2003 |
Jan
|
Feb
|
Mar
|
Apr
|
May
|
Jun
|
Jul
|
Aug
|
Sep
(1) |
Oct
|
Nov
|
Dec
|
|
From: Etienne M. G. <eg...@j-...> - 2000-03-22 18:37:16
|
"Raif S. Naffah" wrote: > i downloaded the UNICODE 3.0 character chart and will proceed to implement > the ECMA specs for the lexer. If you could tell us about all the URLs to related specifications that you find, we could add them to the web page. I personally am interested in knowing the UNICODE 3.0 URL... Etienne -- ---------------------------------------------------------------------- Etienne M. Gagnon, M.Sc. e-mail: eg...@j-... Author of SableCC: http://www.sable.mcgill.ca/sablecc/ ---------------------------------------------------------------------- |
|
From: Raif S. N. <ra...@fl...> - 2000-03-22 18:31:01
|
At 04:06 PM 3/21/00 -0500, Etienne M. Gagnon wrote: >"Raif S. Naffah" wrote: > > > > it may be a SableCC "finesse" that i'm not aware of, but i'd appreciate it > > if somebody can enlighten me on this: > > > > 1. why the (Helpers) definitions for unicode_letter and unicode_digit are > > commented out? > >Because:... >... >The lexer is not done yet. The current definitions were simply "borrowed" >from the Java >grammar. Don't trust any definition. What we know is that there should be a >"white_space" token, but we must find its regular expression. that's what i was hoping to hear. i downloaded the UNICODE 3.0 character chart and will proceed to implement the ECMA specs for the lexer. cheers; rsn |
|
From: Etienne M. G. <eg...@j-...> - 2000-03-21 21:12:33
|
"Raif S. Naffah" wrote: > > it may be a SableCC "finesse" that i'm not aware of, but i'd appreciate it > if somebody can enlighten me on this: > > 1. why the (Helpers) definitions for unicode_letter and unicode_digit are > commented out? Because: (1) These complex regular expressions slow down the lexer generation time. As I was debugging the parser, I commented them out. Quick and dirty;-) (2) These definitions were taken from the "Java" language specification. I am quite unsure if they are appropriate for ECMAScript. So, before you uncomment them, you should doublecheck the specification. I'd like Scriptonite to be 100% conform to the specification. The Parser part already is (but for the literal part). Some digging remains to be done to gather all information about lexer elements. > 2. why do we have > (a) a Helper white_space = (sp | ht | ff | line_terminator)*; and > (b) a Token white_space =; The lexer is not done yet. The current definitions were simply "borrowed" from the Java grammar. Don't trust any definition. What we know is that there should be a "white_space" token, but we must find its regular expression. > 3. the Helper white_space_no_lt = ; is left undefined. any specific reason? See above. > 4. not all UNICODE 3.0 characters are accounted for. is it by design? See above. Etienne -- ---------------------------------------------------------------------- Etienne M. Gagnon, M.Sc. e-mail: eg...@j-... Author of SableCC: http://www.sable.mcgill.ca/sablecc/ ---------------------------------------------------------------------- |
|
From: Raif S. N. <ra...@fl...> - 2000-03-21 20:05:00
|
it may be a SableCC "finesse" that i'm not aware of, but i'd appreciate it if somebody can enlighten me on this: 1. why the (Helpers) definitions for unicode_letter and unicode_digit are commented out? 2. why do we have (a) a Helper white_space = (sp | ht | ff | line_terminator)*; and (b) a Token white_space =; 3. the Helper white_space_no_lt = ; is left undefined. any specific reason? 4. not all UNICODE 3.0 characters are accounted for. is it by design? cheers; rsn |
|
From: Raif S. N. <ra...@fl...> - 2000-03-21 17:13:14
|
At 03:36 PM 3/21/00 +0000, Pedro Vale Lima wrote: >"Raif S. Naffah" wrote: > > > anybody knows if there are suggested/recommended ECMAScript (source) files > > to use for validating the generated lexer and parser? > > > >There is a big test suite in the mozilla project. Look for js-test**** in > >http://ftp.mozilla.org/pub/js/ excellent. i'll go digging... i'll probably create 2 packages under the org.scriptonite hierarchy: org.scriptonite.test and org.scriptonite.test.data the first will contain Test classes, while the second will contain the validation .js test files. >You can also find there rhino, the mozilla project javascript interpreter. >I've just peeked >at it but it seems that the parser was handcoded (?). Anyhow it might be >useful. i'll keep that in mind. thanks for the pointers :-) cheers; rsn |
|
From: Pedro V. L. <eq...@eq...> - 2000-03-21 15:45:39
|
"Raif S. Naffah" wrote: > anybody knows if there are suggested/recommended ECMAScript (source) files > to use for validating the generated lexer and parser? > There is a big test suite in the mozilla project. Look for js-test**** in http://ftp.mozilla.org/pub/js/ You can also find there rhino, the mozilla project javascript interpreter. I've just peeked at it but it seems that the parser was handcoded (?). Anyhow it might be useful. regards pedro |
|
From: Pedro V. L. <eq...@eq...> - 2000-03-21 15:26:43
|
"Etienne M. Gagnon" wrote: > > There's the issue of domain ownership. I can take care of domain > registration, and making arrangements with my ISP to redirect scriptonite.org > to sourceforge, if you decide that I should own the domain. Frankly, I'd like > to know your [all of you] position on this matter. > Although a bit late for the discussion, I just want to say I'm also favorable to Etienne being the keeper of the org.scriptonite domain. It seems Raif's employer will be a sponsor to the registration cost, but if needed I'm also favorable to splitting the cost between the group. regards pedro |
|
From: Raif S. N. <ra...@fl...> - 2000-03-21 09:36:11
|
At 01:41 PM 3/20/00 -0500, Etienne M. Gagnon wrote: >"Raif S. Naffah" wrote: > > > > any objection to generating the java code under the package: > > > > scriptonite > > > > ? > >If any of you wants to pay for registering scriptonite.org, we could use >"org.scriptonite"... i'll add this to the grammar text file and start working on the lexer. cheers; rsn |
|
From: Raif S. N. <ra...@fl...> - 2000-03-21 09:29:41
|
At 06:59 AM 3/21/00 +1100, Raif S. Naffah wrote: >At 02:41 PM 3/20/00 -0500, Etienne M. Gagnon wrote: >>[I had forgotten to CC the mailing-list, and I added some text] >> >>It's 35$US/year. (At least, with Network Solutions. I have had no trouble >>with them. I haven't tried any other registrar.) > >i'll hopefully have an answer this afternoon (down-under time). i'll keep >the group posted. i got a positive answer from Forge management :-) the 35$ (domain registration fee) will be paid by Forge Research Pty. Limited (<http://www.forge.com.au>). Etienne, pls indicate forward address and preferred method to send the money. ...and thank you again for taking this task on board. cheers; rsn |
|
From: Raif S. N. <ra...@fl...> - 2000-03-20 20:04:32
|
At 02:41 PM 3/20/00 -0500, Etienne M. Gagnon wrote: >[I had forgotten to CC the mailing-list, and I added some text] > >It's 35$US/year. (At least, with Network Solutions. I have had no trouble >with them. I haven't tried any other registrar.) i'll hopefully have an answer this afternoon (down-under time). i'll keep the group posted. >There's the issue of domain ownership. I can take care of domain >registration, and making arrangements with my ISP to redirect scriptonite.org >to sourceforge, if you decide that I should own the domain. Frankly, I'd like >to know your [all of you] position on this matter. > >In my opinion, domain costs and domain ownership are separate issues. We can >divide the costs among money contributors (it's quite cheap. If we had 7 >people, 35/7 = 5$/year.) But the ownership should go to somebody who will >take care of keeping the domain alive, and won't take undue advantage of it in >the long term. We probably need concensus on who this person should be. It >doesn't have to be me. if Etienne is willing, i'd like to nominate him as the Keeper. cheers; rsn |
|
From: Etienne M. G. <eg...@j-...> - 2000-03-20 19:47:32
|
[I had forgotten to CC the mailing-list, and I added some text] It's 35$US/year. (At least, with Network Solutions. I have had no trouble with them. I haven't tried any other registrar.) There's the issue of domain ownership. I can take care of domain registration, and making arrangements with my ISP to redirect scriptonite.org to sourceforge, if you decide that I should own the domain. Frankly, I'd like to know your [all of you] position on this matter. In my opinion, domain costs and domain ownership are separate issues. We can divide the costs among money contributors (it's quite cheap. If we had 7 people, 35/7 = 5$/year.) But the ownership should go to somebody who will take care of keeping the domain alive, and won't take undue advantage of it in the long term. We probably need concensus on who this person should be. It doesn't have to be me. Etienne -- ---------------------------------------------------------------------- Etienne M. Gagnon, M.Sc. e-mail: eg...@j-... Author of SableCC: http://www.sable.mcgill.ca/sablecc/ ---------------------------------------------------------------------- |
|
From: Raif S. N. <ra...@fl...> - 2000-03-20 18:55:17
|
At 01:41 PM 3/20/00 -0500, Etienne M. Gagnon wrote: >"Raif S. Naffah" wrote: > > > > any objection to generating the java code under the package: > > > > scriptonite > > > > ? > >If any of you wants to pay for registering scriptonite.org, we could use >"org.scriptonite"... how much money are we talking about? if it's reasonable i might be able to convince my employer to pay [part of] it. cheers; rsn |
|
From: Etienne M. G. <eg...@j-...> - 2000-03-20 18:47:47
|
"Raif S. Naffah" wrote: > > any objection to generating the java code under the package: > > scriptonite > > ? If any of you wants to pay for registering scriptonite.org, we could use "org.scriptonite", and thus conform to the Java convention. Currently, the convention would dictate to use: "net.sourceforge.scriptonite". It is unlikely that anybody will name their project "scriptonite", so one could argue for using package "scriptonite" as you suggest. I personally have no objection to any of these three options. -- ---------------------------------------------------------------------- Etienne M. Gagnon, M.Sc. e-mail: eg...@j-... Author of SableCC: http://www.sable.mcgill.ca/sablecc/ ---------------------------------------------------------------------- |
|
From: Raif S. N. <ra...@fl...> - 2000-03-20 18:41:52
|
any objection to generating the java code under the package: scriptonite ? cheers; rsn |
|
From: Raif S. N. <ra...@fl...> - 2000-03-20 18:02:34
|
anybody knows if there are suggested/recommended ECMAScript (source) files to use for validating the generated lexer and parser? cheers; rsn |
|
From: Pedro V. L. <eq...@eq...> - 2000-03-09 10:58:27
|
"Etienne M. Gagnon" wrote: > Now, it's your turn to start contributing! Don't wait for me to write this > project; I won't. But I am willing to answer questions and offer guidance. No answer so far.... humm. I'm just speculating, but I think the people in the list are just beginning to learn SableCC (at least I am). Unless someone really needs this project, I guess we will have to be happy with a slow development. But I don't expect you (Etienne) to write the all thing. This is to have fun (for pain we all have our real works right?). Although with slow progress I'm still with you gang! The next step is building the custom lexer so we can parse some simple scripts (right?). Have Fun! pedro |
|
From: Etienne M. G. <eg...@j-...> - 2000-03-05 21:18:57
|
Hi!
I have used the latest SableCC release to solve all conflicts in the grammar.
Notice that a few irregulars requirements of the specification will require
the help of a custom lexer.
Now, it's your turn to start contributing! Don't wait for me to write this
project; I won't. But I am willing to answer questions and offer guidance.
Here is a summary of the conflicts/solutions:
(1) no line terminator in front of incr/decr (++/--): Define the two versions
of tokens that include preceeding white space with or without line
terminators. Fix the grammar appropriately.
(2) no line terminators after continue, break, return, throw. Solved with
lexer states and custom lexer. No grammar modifications.
(3) statement_expression should be an expression that does not start with "{"
or "function". Unroll grammar (introducing expression_no_lbf).
(4) dangling-else ambiguity. Unroll grammar (statement_no_if).
This was it. I guess that the designers of ECMAScript used an LR parser;-)
Etienne
|
|
From: Etienne M. G. <eg...@j-...> - 2000-03-03 17:50:41
|
Hi. Has anybody made any progress, so far? Etienne |
|
From: Uwe H. <uw...@ia...> - 2000-02-28 22:21:13
|
> > What was the problem? I mean, what caused it to fail on Solaris, > previously? > I'm not sure. But I was wrong. It was not the same error message. It had something to do with my ssh setup. I had to delete my .ssh and log in again with ssh. After ssh worked cvs worked also. |
|
From: Etienne M. G. <eg...@j-...> - 2000-02-28 21:28:18
|
What was the problem? I mean, what caused it to fail on Solaris, previously? Etienne |
|
From: Uwe H. <uw...@ia...> - 2000-02-28 21:07:38
|
I got it working on solaris. I'll work from there. thanks guys |
|
From: Uwe H. <uw...@ia...> - 2000-02-28 20:50:44
|
> When you used "ssh", did you need to provide any option? e.g.: > > ssh -P -l uweDeportivo cvs.scriptonite.sourceforge.net ssh -l uwedeportivo cvs.scriptonite.sourceforge.net that worked > > [ > By the way: The line above shoud result into: > Last login: xxx from xxx > No mail. > > Welcome to cvs1.sourceforge.net > > This is a restricted Shell Account > You cannot execute anything here. > > Connection to cvs.scriptonite.sourceforge.net closed. > ] > > If this works, then you need to make a small batch file called > MySSH.BAT with > the content: > ssh -P -l uweDeportivo %1 %2 %3 %4 %5 > > and you set CVS_RSH to MySSH (which should also be on your path). > > Then try: > MySSH cvs.scriptonite.sourceforge.net that worked too > > If this works (e.g. same result as above), then try cvs -d again. > > I hope this works... unfortunately I still get cvs [checkout aborted]: cannot connect to socket: Connection refused |
|
From: Raif S. N. <ra...@fl...> - 2000-02-28 19:59:53
|
At 10:56 AM 2/28/00 -0800, Uwe Hoffmann wrote: >I'm still in setup hell on nt.... have a look at <http://www.cryptix.org/docs/devcvs.html>. we use cvs too and i wrote this page to help NT developers use NT-based tools. cheers; rsn |
|
From: Etienne M. G. <eg...@j-...> - 2000-02-28 19:29:42
|
Hi Uwe. When you used "ssh", did you need to provide any option? e.g.: ssh -P -l uweDeportivo cvs.scriptonite.sourceforge.net [ By the way: The line above shoud result into: Last login: xxx from xxx No mail. Welcome to cvs1.sourceforge.net This is a restricted Shell Account You cannot execute anything here. Connection to cvs.scriptonite.sourceforge.net closed. ] If this works, then you need to make a small batch file called MySSH.BAT with the content: ssh -P -l uweDeportivo %1 %2 %3 %4 %5 and you set CVS_RSH to MySSH (which should also be on your path). Then try: MySSH cvs.scriptonite.sourceforge.net If this works (e.g. same result as above), then try cvs -d again. I hope this works... Etienne Uwe Hoffmann wrote: > > I'm still in setup hell on nt. > > If one of you has an idea about the following it would be great. > > What I have so far: > > - installed all required pieces (cvs, ssh) > - logged in with ssh: worked > - checked out with cvs and anonymous pserver: worked > - tried to check out with CVS_RSH=ssh and -d username argument: didn't work > and got > > cvs [checkout aborted]: cannot connect to socket: Connection refused > > - I thought maybe it's some of my webservers: killed all apache and jrun > instances: still the same > > - went to our solaris machine and tried there: same problem > > Is it possible that the cvs server side doesn't let me in ? > > thanks, > > uwe > > _______________________________________________ > Scriptonite-dev mailing list > Scr...@li... > http://lists.sourceforge.net/mailman/listinfo/scriptonite-dev -- ---------------------------------------------------------------------- Etienne M. Gagnon, M.Sc. e-mail: eg...@j-... Author of SableCC: http://www.sable.mcgill.ca/sablecc/ ---------------------------------------------------------------------- |
|
From: Uwe H. <uw...@ia...> - 2000-02-28 18:58:29
|
I'm still in setup hell on nt. If one of you has an idea about the following it would be great. What I have so far: - installed all required pieces (cvs, ssh) - logged in with ssh: worked - checked out with cvs and anonymous pserver: worked - tried to check out with CVS_RSH=ssh and -d username argument: didn't work and got cvs [checkout aborted]: cannot connect to socket: Connection refused - I thought maybe it's some of my webservers: killed all apache and jrun instances: still the same - went to our solaris machine and tried there: same problem Is it possible that the cvs server side doesn't let me in ? thanks, uwe |