ticalc.org
Basics Archives Community Services Programming
Hardware Help About Search Your Account
   Home :: Archives :: News :: TiLP's Final Release

TiLP's Final Release
Posted by Michael on 4 July 2009, 20:47 GMT

Romain Liévin has announced that he is discontinuing development of his immensely popular TiLP emulator software. After ten years of 68k calculator development, Romain believes that it is time to move on to other activities. The recent decline of activity in the TI calculator community and complexity of his projects (TiLP, TiEmu, and GFM) contributed to his retirement decision. We will miss Romain and wish him luck in his other projects.

  Reply to this article


The comments below are written by ticalc.org visitors. Their views are not necessarily those of ticalc.org, and ticalc.org takes no responsibility for their content.


Re: TiLP's Final Release
Kevin Kofler Account Info
(Web Page)

Development of TiLP and its libraries and of TiEmu will continue (under new names at Romain's request - start getting used to CalcForgeLP and Emu-TIGCC) at http://www.calcforge.org/ . I (coauthor of TiEmu and to some extent of TiLP and the libraries) will be the main developer, Tyler Cassidy (the original author of the Group File Manager) will be the administrator and co-developer. Right now we're working on removing unneeded cruft and changing the names everywhere so we can do a first release.

Reply to this comment    4 July 2009, 20:55 GMT

Re: Re: TiLP's Final Release
Kevin Kofler Account Info
(Web Page)

PS: The full list of new names:
tfdocgen -> CalcForge docgen
libticables2 -> libcalccables
libticalcs2 -> libcalcprotocols
libticonv -> libcalcconv
libtifiles2 -> libcalcfiles
TiLP 2 -> CalcForgeLP
GFM -> CalcForgeLP GFM
TiEmu -> Emu-TIGCC
SkinEdit -> Emu-TIGCC SkinEdit

Our plans (short-term):
* port of the KDE integration from KDE 3 to KDE 4
* support for the asynchronous API of libusb 1.0 (Emu-TIGCC needs asynchronous transfers for the SilverLink support, to detect whether data is ready on the USB port. There’s currently an implementation using the asynchronous API of libusb-win32, which shall remain present, and a special hack for Linux using internal platform-specific libusb structures. Using libusb 1.0 should obviate the need for this hack and allow supporting the SilverLink in Emu-TIGCC also on other *nix systems.)
* migration of the build system from the autotools to CMake
* replacement of private copies of libraries by system versions: minizip in libcalcfiles and Tcl/Tk/itcl/itk/iwidgets in Emu-TIGCC shall be replaced by system versions, at least where those are available
* removal/deactivation of the UAE disassembler in Emu-TIGCC in favor of the libopcodes/GDB disassembler (which can also be used without GDB, see e.g. dasm-tigcc)
* removal of Visual C++ compiler support (MinGW and cross-MinGW will remain supported!)
Long-term, we also do not exclude extending CalcForgeLP and the corresponding libraries to also support calculators by other manufacturers.

Reply to this comment    4 July 2009, 21:05 GMT

Re: Re: TILP's Final Release...
Lionel Debroux Account Info
(Web Page)

Mod parent -1 lie.
Lies call for facts being faced against them, so here we go...

Kevin, it's obvious that you're eager to promote TILP-fork, TIEmu-fork, libti*-forks and the rest of the LPG software that you've recently forked.
However, it's dishonest of you to try making people believe that your forks are the new upstream ("Development of TiLP and its libraries and of TiEmu will continue ..."). It ain't so.
The FACT is: the development of TILP, of its libraries and of TIEmu, will continue at the usual place, http://lpg.ticalc.org/ .

It's dishonest of you not to mention that you could still be a member of TIEmu and the others, if you hadn't behaved in such an unacceptable way at [see URL below poster name] (Romain's announcement, posted several days after the news was sent to ticalc.org, i.e. several weeks ago). In that topic, it's a FACT you did:
* auto-proclamation of maintainership without having discussed with the maintainers;
* posting of an explicit "I am the maintainer, I can do what I want", and other evidences of disregarding community members' opinion & flat out refusing collaborations that don't match your goals (Romain and Julien have always led the project in a much more open-minded way, so that change would definitely have been unwelcome for the community);
* minimization and criticism of the work done by Romain and Julien.
Well, that's not how collaborative, community-minded development of FLOSS works.

Reply to this comment    5 July 2009, 07:36 GMT

Re: Re: Re: TILP's Final Release...
Travis Evans Account Info

*Digs out fire extinguisher* Can never be too prepared around fireworks. ;-)

Reply to this comment    5 July 2009, 10:26 GMT


Re: Re: Re: Re: TILP's Final Release...
Lionel Debroux Account Info
(Web Page)

;-)
Don't worry, if Kevin doesn't pile more crap over his crap I've replied to, it's unlikely that there will be additional fire in this topic.

We'd MUCH rather focus on doing actual work than on having to spend time explaining the facts that were twisted by Kevin... It's a REAL shame he can't seem to do without attacking other people and other people's work.
If nothing else, he'd better backport to TIGCC bugfixes that the GCC4TI project made and reported weeks ago, or optimizations and documentation updates that were contributed to TIGCC as early as 2002 (!) and whose double-checking and integration by GCC4TI has begun...

Reply to this comment    5 July 2009, 11:10 GMT


Re: Re: Re: Re: Re: TILP's Final Release...
Kevin Ouellet Account Info
(Web Page)

I agree, I think it's best if you contribute to the community instead of arguing over and over. It's always similar arguments that pop up anyway and people are getting rather tired of them. In fact, I think this is one of the reason (not mentioned in the Ticalc.org news article) why the original author is retiring.

Plus, since it's the internet, nothing stops you from working on whatever calc projects you want. Unless you violate GPL licenses and author rights, the other person can do nothing about it. All he can do is argue and flame, in which case you can simply filter his e-mails/messages and posts

Reply to this comment    5 July 2009, 21:31 GMT


Re: Re: Re: TILP's Final Release...
Kevin Kofler Account Info
(Web Page)

You just have to follow my link (http://www.calcforge.org/) to read that there's your attempt to continue LPG. So I'm not denying that it exists. But the FACT is that the CalcForge project is maintained by the remaining existing LPG contributors (Tyler and me - Tyler started the GFM and you just have to look at the changelogs to see how much work I did: integration of GDB/Insight into TiEmu (alone, Romain refused to help because he said that code is too complex for him), sound emulation in TiEmu together with Peter Fernandes (hypersonic), KDE integration (TiEmu, TiLP, GFM), many small patches (improvements, bugfixes etc.) throughout all the LPG software) whereas you are the new guy and the 2 friends you're "working with" are both inactive (Romain just left and Julien has abandoned LPG development and been doing only administrative tasks for years). So I consider your effort to be the fork.

Reply to this comment    5 July 2009, 12:07 GMT


Re: Re: Re: Re: TILP's Final Release...
Lionel Debroux Account Info
(Web Page)

What matters in this discussion, is the FACT that you didn't mention the LPG effort here, in this discussion. Nor you mentioned that your effort is a fork. You're twisting the truth.
We GCC4TI contributors (seven people participated to the latest release) don't try painting GCC4TI as the official continuation of TIGCC - though GCC4TI has less bugs and more optimizations than TIGCC has, and you have done almost nothing on TIGCC since January (first GCC4TI release).

Those who decide which side is a fork, are the maintainers of the original tree. And it's very clear which side is a fork: yours.
Therefore, please quit propagating the LIE that the LPG continuation of those programs is the fork. This is a LIE, and repeating it because you're unconvinced won't make it magically become truth. Quit being a LIAR, period.

If I were you, I'd be worried that despite all your contributions to TIEmu, Romain and Julien didn't automatically, upon announcement of Romain's returiement, hand over that project's maintenance to you...
They'd have done it if they trusted you to advance the project in the best direction for the community, but they don't... with reason, since for example, you want to discard the '%'-free syntax without letting users the CHOICE, nor asking them first for input.

I'm not a newcomer in TIEmu: I've contributed bug reports, feature requests, and tests for about as long as you have.

Reply to this comment    5 July 2009, 13:05 GMT

Re: Re: Re: Re: Re: TILP's Final Release...
Kevin Kofler Account Info
(Web Page)

> Those who decide which side is a fork, are the maintainers of the original tree.

And that's what I've been before I got kicked out of the infrastructure. I've been codeveloping/comaintaining TiEmu and TiLP for years. Romain coordinated releases with me (at least most of the time), I've done a lot of development. It's clear that I was the developer #2 in the project and that the #1 was Romain who just left. Draw your own conclusions.

See, "the original tree" is also a subjective term. I disagree with the assertion that the "original tree" is the one sitting on the original infrastructure. The infrastructure people are using is actually irrelevant, what's relevant is the people working on it.

Your comparison with GCC4TI is flawed: the continuing TIGCC isn't just using the existing infrastructure, it's developed by the remaining active TIGCC Team member (me). Just like CalcForgeLP and Emu-TIGCC are being developed by the remaining active members of what used to be LPG (at least the TiLP/TiEmu/tilibs portion - TilEm is a different story, it used to be completely separate before Romain had it imported into svn.tilp.info and it is being continued in only one place now). So it's CalcForgeLP and Emu-TIGCC which are like TIGCC and the "new LPG" which is like GCC4TI.

Reply to this comment    5 July 2009, 13:22 GMT


Re: Re: Re: Re: Re: TILP's Final Release...
Kevin Kofler Account Info
(Web Page)

> If I were you, I'd be worried that despite all your contributions to TIEmu, Romain and Julien didn't automatically, upon announcement of Romain's returiement, hand over that project's maintenance to you...

I'm not worried, I'm offended and I blame all 3 of you for it.

> I'm not a newcomer in TIEmu: I've contributed bug reports, feature requests, and tests for about as long as you have.

That's something a user does, not a developer. Your contributions as a developer are limited to a small number of simple patches which (if I'm not mistaken) can be counted on the fingers of one hand.

Reply to this comment    5 July 2009, 13:22 GMT


Re: Re: Re: Re: Re: Re: TILP's Final Release...
Lionel Debroux Account Info
(Web Page)

The conclusion that Romain, Julien and many others had drawn, long before Romain stopped maintenance on the LPG programs, is that you suck as #1.
You made a great show of suckiness as a leader project in [see URL above post]: disregarding other peoples' input, minimizing work done by other maintainers, explicitly writing "I am the maintainer, I can do whatever I want", etc.

The other good example of your suckiness as a leader, is TIGCC:
* even contributions that you deem nice, and even those which are easy to review and test, are delayed for years. Meanwhile, you spend hours and hours trolling about proprietary software, dissing other peoples' work, etc.;
* many contributions pushed by multiple persons, using technical arguments, have been unilaterally rejected;
* you have a priority ordering different from that of many TIGCC users;
* TIGCC's quality sucks. The absence of a good automated build system is one thing, but I was astounded when I noticed that merely trying to build and run the examples ("useless", as you wrote on yAronet) caught:
* three tool bugs (one of which you already knew about, but most of the community didn't - I independently rediscovered it !);
* a compilation failure present since 2005 (which shows you haven't even used the examples as regression tests & benchmarks for the GCC versions released since then);
* two warnings;
* a couple dozen name clashes between examples (that is on purpose, you wrote on yAronet).

Blame us if you wish, and blame people for giving up trying to work with you in TIGCC. However, it would be more useful to question yourself about your behaviour - and most of all, fix that destructive behaviour of yours...

I'm a "user" of TIEmu who has, in 2009, contributed or committed more patches to TIEmu than you have.

Reply to this comment    5 July 2009, 14:12 GMT


Re: Re: Re: Re: Re: Re: Re: TILP's Final Release...
Kevin Kofler Account Info
(Web Page)

I didn't write that the name clashes are "on purpose", I wrote that they're not worth fixing as the examples aren't supposed to be sent to the calculator at the same time anyway. You seem to misunderstand the purpose of the examples completely. The examples are there as a reference. They should be READ and the insights gained incorporated in actual programs. They are NOT intended to be used as is. In particular, the examples are NOT benchmarks, they're intentionally NOT optimized, but written for maximum readability. (In fact, I rejected a patch you sent once to optimize an example at the expense of readability.) They should be treated as documentation, NOT software.

Reply to this comment    5 July 2009, 15:13 GMT


Re: Re: Re: Re: Re: Re: Re: Re: TILP's Final Release...
Lionel Debroux Account Info
(Web Page)

> You seem to misunderstand the purpose of the examples completely.
Not at all, I'm just using them for a slightly different purpose. And it turned out to be pretty useful to do that !
Regression testing is a very basic good practice of software engineering, so adding regression testing wasn't even a proof of creativity on my side. It was, however, a proof that the TIGCC quality standards, inherited by GCC4TI, are very low...

> They are NOT intended to be used as is.
Well, I don't know for you, but when _I_ see code examples that are ready for use, I execute them. That's what I did recently with the Qt examples bundled with Qt Creator.
If that means compiling the examples first, so be it.

> the examples are NOT benchmarks,
They CAN be used as such (size benchmarks, of course), when testing a new version of one of the tools that generate or consume object code.
> they're intentionally NOT optimized
I know that.

> They should be treated as documentation, NOT software.
That's YOUR opinion. We beg to differ ;-)
The fact is, the 60+ examples exercise various areas of the TIGCC library and the toolchain (e.g. floating-point support). As such, they CAN be used as an early warning system for breakage. If even those trivial testcases have serious problems, it's of little use pushing new tool / TIGCCLIB versions to users.

Reply to this comment    5 July 2009, 16:01 GMT


Re: Re: TiLP's Final Release
Lionel Debroux Account Info
(Web Page)

It's dishonest of you not to mention that while you changed your forks' names, at first, it's a FACT that you refused to do so, pretexting that you were the real upstream and that the others were just staging a "coup d'état". You changed your mind only after Romain complained very hard about such a disgusting misappropriation and disrespect of the software he'd led for so many years.

People who work on TILP, libti* and TIEmu don't need to spend hours performing the nonfunctional change that changing names is. They don't feel the urge for e.g. dropping documentation, dropping MSVC support (what you term "unneeded cruft").

I wouldn't have posted these two comments if Kevin hadn't posted lies in the first place. Just as he did for the GCC4TI news...

Reply to this comment    5 July 2009, 07:45 GMT


Re: Re: Re: TiLP's Final Release
Kevin Kofler Account Info
(Web Page)

I didn't want to change the names at first because it's a lot of work (we finally completed the first of the 9 renames and that was the easiest one) and because except for Romain (who's leaving) we're the same people who always worked on it. We ended up changing them anyway 1. because I don't want to create bad blood with Romain over something as trivial and 2. to avoid user confusion between your and our projects.

Reply to this comment    5 July 2009, 12:18 GMT


Re: Re: Re: Re: TiLP's Final Release
Lionel Debroux Account Info
(Web Page)

Yeah, renaming things is a time-consuming, nonfunctional change, which brings nothing to users...

You would have avoided yourself all the pain, and simultaneously advanced the software sooner, if you had been more reasonable in your relationship with the other maintainers and with the programs' users...

The other maintainers DO think that unifying the TIEmu disassemblers is a good thing. We've told that multiple times.
However, we side with users on that imposing on everybody (without asking first for input) the lowercase, '%'-littered, decimal+octal GDB syntax, instead of leaving the CHOICE between your preferred syntax and the uppercase, '%'-less, hexadecimal VTI syntax that many prefer, is not good.

Reply to this comment    5 July 2009, 13:11 GMT

Re: Re: Re: Re: Re: TiLP's Final Release
Kevin Kofler Account Info
(Web Page)

It's a lot of work to retrofit the legacy syntax into the libopcodes disassembler (it's especially hard to support both syntaxes at once in the same disassembler, it needs conditionals everywhere, and no, post-processing is not a solution, we've been through that discussion already) and it delays the cleanup, for no concrete benefit, as the syntax the disassembler outputs makes no functional difference.

Reply to this comment    5 July 2009, 13:24 GMT


Re: Re: Re: Re: Re: TiLP's Final Release
Kevin Kofler Account Info
(Web Page)

Oh, and I can definitely add an option to show all numbers in hex (that's a useful feature). But the hex will be prefixed with "0x" (as in GNU as), not '$'. (That reminds me that I need to fix all the '$' signs in Emu-TIGCC to 0x.)

Reply to this comment    5 July 2009, 13:27 GMT


Re: Re: Re: Re: Re: Re: TiLP's Final Release
Kevin Kofler Account Info
(Web Page)

PS: Octal is only used for invalid instructions.

Reply to this comment    5 July 2009, 13:37 GMT

Re: TiLP's Final Release
Lionel Debroux Account Info
(Web Page)

TILP II 1.13, TIEmu 3.03 and the latest current versions of the associated libraries were the last ones released by Romain, but that does not mean that the projects are dead ;-)

What would happen after Romain's retirement was intentionally not mentioned in the news, so as to see whether some people would step in to help. Someone (no, not Kevin Kofler) did on yAronet, and he started working on TILem and libti*. However, TILP, TIEmu and the associated projects are looking for more people.

Reply to this comment    5 July 2009, 07:21 GMT


Re: Re: TiLP's Final Release
Kevin Kofler Account Info
(Web Page)

> Someone (no, not Kevin Kofler) did on yAronet, and he started working on TILem and libti*.

If you want to work on TilEm, I recommend to just work with that "someone" (his nickname on yAronet is "Contra"). We will not be maintaining TilEm on CalcForge (except for RPM packaging), because it's not Free Software (the Z80 emulation core is under a license forbidding commercial use) and because none of us was actively involved in TilEm development even before.

> However, TILP, TIEmu and the associated projects are looking for more people.

As are CalcForgeLP, Emu-TIGCC and the associated projects. If you're interested, come to our IRC chan (#tigcc): see http://tigcc.ticalc.org/webchat.html for the precise address, channel rules and a CGI webchat.

Reply to this comment    5 July 2009, 12:33 GMT


Re: Re: Re: TiLP's Final Release
Kevin Kofler Account Info
(Web Page)

PS: You'll also find the TilEm guy ("Contra") on #tigcc, his nickname on IRC is "azerti".

Reply to this comment    5 July 2009, 12:44 GMT

  Copyright © 1996-2011, the ticalc.org project. All rights reserved. | Contact Us | Disclaimer