From: Borut R. <bor...@si...> - 2010-01-13 19:31:29
|
Hello sdcc users and developers, here is a list of tasks which by my opinion are candidates for the sdcc 3.0 release: * sdas merge with asxxxx 5.0 * change sdcc libraries license to GPL+LE * review and merge "SDCC inline support and fixes" enhancements by Pavel Pisa see http://sourceforge.net/tracker/?func=detail&atid=100599&aid=1767885&group_id=599 * review and merge support for cs08 target by Gary Osborn * deprecate warnings for sdcc specific keywords, not preceded with two underlines (data, code, idata, xdata, ...) and announce in the documentation that they will be removed in the future * remove unsupported and broken targets xa51, avr * review & merge patches in the bug tracker * bug fixing * ... The list is quite long and you probably have additional candidates. I propose to make a list of candidates on sdcc wiki, prioritize them and find the persons which will work on them. There are many tasks that can be done by non "core" sdcc developers: change sdcc libraries license, changes, review and reorganization of the documentation, ... So if there is anybody willing to help us, let us know! Borut |
From: Borut R. <bor...@si...> - 2010-01-15 08:00:40
|
Now I have 2 skeptic/conservative opinions from sdcc developers and 2 optimistic replays from sdcc users. I would like to have more opinions from sdcc users, so don't hesitate to send a mail to sdcc-user or sdcc-devel mailing list. In my previous post I asked sdcc users to help us (sdcc developers) in some tasks, mainly in: * change sdcc libraries license to GPL+LE: collect the info and make a list of sdcc library files for which the license can be changed, based on the agreement from authors (see http://sdcc.sourceforge.net/wiki/index.php?page=Library+License+Selection) * documentation upgrade / review / reorganization * ... Anybody willing to take the challenge? Borut Sébastien Lorquet wrote: > Agree too, for the same reasons :) > Sebastien > > On Thu, Jan 14, 2010 at 11:12 PM, steven.borley > <ste...@vi... <mailto:ste...@vi...>> wrote: > > Borut, > > As a SDCC user I'd have no problem with the changes you propose. > > As for the deprecation of the keywords I long ago switch to using > only the version that begin with the double underscore. > > One consequence of this would perhaps be to make porting code > easier. Form experience I often find 'data' is used as a variable > name, which currently causes some very odd error messages to an > unsuspecting user. > > Steven > > > On 14 Jan 2010, at 06:54, Borut Razem wrote: > > > Hello Jesus, > > > > here is the complete list of deprecated keywords: > > > > non-compliant | compliant > > keywords | keywords > > ===============+============= > > at | __at > > bit | __bit > > code | __code > > critical | __critical > > data | __data > > far | __far > > eeprom | __eeprom > > fixed16x16 | __fixed16x16 > > flash | __flash > > idata | __idata > > interrupt | __interrupt > > nonbanked | __nonbanked > > banked | __banked > > near | __near > > pdata | __pdata > > reentrant | __reentrant > > shadowregs | __shadowregs > > sfr | __sfr > > sfr16 | __sfr16 > > sfr32 | __sfr32 > > sbit | __sbit > > sram | __sram > > using | __using > > _naked | __naked > > xdata | __xdata > > _overlay | __overlay > > > >> I believe this will create a whole bunch of compatibility issues. > > > > It will produce a bunch of warnings if the code use > non-compliant keywords. > > The code will still compile correctly. And if somebody really > don't want > > to see the wraninings, he can use "#pragma disable_warning 197" > in the > > source file or "--disable-warning 197" command line option. > > > > I actually already implemented this functionality and it was > trivial: > > just introduce the error message in SDCCerr.[ch] and change the > > TKEYWORDSDCC macro in SDCC.lex, so it can be easily reverted if > there > > are too many complains from sdcc users. > > > > The harder work was correcting the regression tests, which used > > non-compliant keywords a lot. > > > > Best regards, > > Borut > > > > > > > > Jesus Calvino-Fraga wrote: > >> Hi Borut, > >> > >> Can you explain why sdcc specific keywords need to be > deprecated? I > >> believe this will create a whole bunch of compatibility issues. > >> > >> Jesus > >> > >> At 11:31 AM 13/01/2010, you wrote: > >> > >>> Hello sdcc users and developers, > >>> > >>> here is a list of tasks which by my opinion are candidates for > the sdcc > >>> 3.0 release: > >>> > >>> * sdas merge with asxxxx 5.0 > >>> * change sdcc libraries license to GPL+LE > >>> * review and merge "SDCC inline support and fixes" > enhancements by > >>> Pavel Pisa > >>> see > >>> > >>> > http://sourceforge.net/tracker/?func=detail&atid=100599&aid=1767885&group_id=599 > <http://sourceforge.net/tracker/?func=detail&atid=100599&aid=1767885&group_id=599> > >>> * review and merge support for cs08 target by Gary Osborn > >>> * deprecate warnings for sdcc specific keywords, not > preceded with > >>> two underlines (data, code, idata, xdata, ...) and > announce in the > >>> documentation that they will be removed in the future > >>> * remove unsupported and broken targets xa51, avr > >>> * review & merge patches in the bug tracker > >>> * bug fixing > >>> * ... > >>> > >>> The list is quite long and you probably have additional > candidates. > >>> > >>> I propose to make a list of candidates on sdcc wiki, > prioritize them and > >>> find the persons which will work on them. There are many tasks > that can > >>> be done by non "core" sdcc developers: change sdcc libraries > license, > >>> changes, review and reorganization of the documentation, ... > >>> > >>> So if there is anybody willing to help us, let us know! > >>> > >>> Borut > >>> > |
From: Gordon H. <gor...@dr...> - 2010-01-15 09:17:32
|
I think I've deleted/lost the start of this thread, however as a brand new sdcc-user (well 2-3 months) one thing I'd like to see is eradication of a lot of the old articles about how poor the SDCC PIC support is... I'm nearing completion of my project - it's about 11K lines of C and compiled into about 90KB of PIC18 code and I really don't know what I'd have done without SDCC (ok, i'd have paid for a commercial compiler, but that's not the point!) It's not been without it's issues - executing math code in interrupts, for which I made a workaround in my RTOS for - however one of the commercial vendus I approached doesn't support that either... Another thing I've read recently is that the sparkfun people (Arduino) have been approached by the Microchip people to make a PIC version of Arduino - and the usual complaints follow of no good C compiler.... (Although they use c++ which isn't for me, but there might still be milage there) So maybe there's scope for more documentation and maybe a "press release" or 2? But I really wouldn't know where to start myself... Actually, I can - when I have time, I'm going to create a new blog for my project when it's finally ready, so will push SDCC there. It's a real-time dive computer if anyones interested - based on the OSTC Mk 2. (Which is "open source" hardware) I'm not sure yet if I'm going to sell it or give it away - I've not decided that yet. Gordon |
From: Borut R. <bor...@si...> - 2010-01-15 09:02:14
|
Sébastien Lorquet wrote: > I guess the steps would be like this: > > -list all lib files > -stuff an excel/scalc sheet > -get their current licence > -get their author > ... > -profit. > Exactly. We already have some tables on sdcc wiki, but this is not applicable for such big number of entries. I vote for scalc. We might even create a database on sourceforge and feed the web page from it, but I think the overhead is to high. If someone wants to play with it, welcome! > I will try to start this. The first 2 steps are quick, but... my > explorer says: in lib/src, there are 666 files and 40 folders. > > Wow. > Great, thanks! Maybe it would make sense to divide the task on per target base so that many people can work in parallel in case we will have additional volunteers. > This will take time :) Yes, that's why I'm looking for help ;-) > What is the estimated timeframe to complete this? > Currently my schedule for the sdcc 3.0 release is end of march (1 year after 2.9 release). But I don't mid if this will be shifted for a month or so since we have some great and huge tasks to do. > Regards > Sebastien. > Borut |
From: Philipp K. K. <pk...@sp...> - 2010-01-15 09:11:46
|
Borut Razem schrieb: > * deprecate warnings for sdcc specific keywords, not preceded with > two underlines (data, code, idata, xdata, ...) and announce in the > documentation that they will be removed in the future I don't know about the other targets, but at least for Z80 those keywords are more trouble than use, mostly they have no real purpose, thus are trouble only and of no use. A warning is a step in the right direction; it will help users that use these keywords fe.g. for naming variables diagnose their problem. I hope we can remove them in the next release after 3.0. Philipp |