Index
> Main > ? :Goto page Previous 1, 2, 3 Next |
| Author |
|
|
revolution 28 Nov 2010, 12:27
Tomasz Grysztar wrote: It is just a number that is generated at the time when the file is uploaded there. That this number is at the same time some readable date is just an additional nice feature. |
|||
28 Nov 2010, 12:27 |
|
|
Tomasz Grysztar 28 Nov 2010, 12:30
I guess no one ever does that, because it not really needed for anything. If you see that version or data incremented in an obvious way, you just grab the package, otherwise quite possibly the change is so minor that it has no immediate consequence for you (unless you are the one that reported something - they you get notified in a reply).
|
|||
|
28 Nov 2010, 12:30 |
|
|
ouadji 28 Nov 2010, 12:41
Quote:
(my English is too poor to correctly explain.) once again, no matter if the update is major, minor ... or very very very small ... a silent update is a invisible update, here is the problem Quote: so I may as well not do that in the future. _________________ I am not young enough to know everything (Oscar Wilde)-  |
|||
|
28 Nov 2010, 12:41 |
|
|
JohnFound 28 Nov 2010, 13:14
Come on, guys! IMHO, this theme it is totally not important. Even if you miss some "silent" update, what of it? Nothing bad will happen. You will get these features in the next major update. It is not as you are writing software for the space industry.
On the other hand, the one that really needs these updates, because he stuck on some bug that stops his work - well this one usually knows about the update, because the update is a result of his request. |
|||
|
28 Nov 2010, 13:14 |
|
|
ouadji 28 Nov 2010, 14:41
Quote: It is not as you are writing software for the space industry ... and much more complex  
_________________ I am not young enough to know everything (Oscar Wilde)- |
|||
|
28 Nov 2010, 14:41 |
|
|
killasmurf86 28 Nov 2010, 17:18
Much have been said...
I just wanted to explain why I asked, not to do that again.... FreeBSD is auditing FreeBSD ports for potential problems, like broken links, mismatched checksums, availability of mirrors etc. This morning I recieved email, that checksum stored with FreeBSD fasm port didn't match archive available on flatassembler.net... What this means? This means that someone can't install fasm, because depending on from where archive was fetched (FreeBSD mirrors or fasm homepage), archive checksum doesn't match. It is not a big deal if checksum doesn't match for freebsd mirros, because in this case file will be download from MASTERSITES (in this case flatassembler.net). but when checksum doesn't match for file from MASTERSITES, port is pretty much broken.... On one hand it might not be big deal, because you just update port in about 5-10 minutes and submin Problem Report... However problem is that It takes time before someone will commit your patch, because there are may PRs waiting to be reviewed, checked, tested, fixed and committed. Sometimes it takes day, sometimes it can take month
if I may suggest... adding additional digit for these small updates: fasm-1.69.27 fasm-1.69.27.1 Everyone would know that it's very minor update, that for most of us doesn't matter, but for automated tools, it would help avoid problems
Anyway... Thanks for you precious time Tomasz
FASM rocks EDIT: I will take note, and update fasm port after some day or 2 in future |
|||
|
28 Nov 2010, 17:18 |
|
|
Tomasz Grysztar 28 Nov 2010, 17:26
Don't they have any option to disable checksum checking for a specific download? I think they should have some possibility to accommodate to a different packaging philosophies.
|
|||
|
28 Nov 2010, 17:26 |
|
|
killasmurf86 28 Nov 2010, 17:29
AFAIK no, and that would name port vulnerable...
Imagine if someone hacks flatassembler.net and adds some harmful code 
Once you fail to download from all MASTERSITES specified (in this case only 1), port won't compile.... User can remove distinfo and update port manually etc... but that is out of scope... Checksums are there for a reason |
|||
|
28 Nov 2010, 17:29 |
|
|
revolution 28 Nov 2010, 17:47
Well so far it seems all commentors in this thread are people that use fasm to make other tools. I think this says a lot about the motives for why we all (except JohnFound) request unique version numbering - so that we can update our tools and keep up-to-date with any bugs or problems (although I am surprised that JohnFound seems not to care that Fresh might be out-of-date). Anyhow, I guess Tomasz has his reasons (although so far he has not stated them) for wanting to avoid using up excessive amounts of new version numbers, so rather than pester him more I will just accept that he wants to keep the IDE and macro updates as silent.
JohnFound: Does Fresh make use of the fasm macros like stdcall/resource/etc. from the download? Or do you provide you own custom versions of these macros? I ask because these "uncore" updates to macros might affect Fresh users also. |
|||
|
28 Nov 2010, 17:47 |
|
|
Tomasz Grysztar 28 Nov 2010, 18:31
killasmurf86 wrote: Checksums are there for a reason Because assuming only the release-based approach seems to me like forcing the one and only one valid distribution philosophy. Hmmm, or maybe I should stay quiet, as they will soon release some EU regulations about how software should be distributed and I will have to toe the line.  |
|||
|
28 Nov 2010, 18:31 |
|
|
killasmurf86 28 Nov 2010, 18:49
Well, you can make fasm-current (latest and greatest)
|
|||
|
28 Nov 2010, 18:49 |
|
|
JohnFound 28 Nov 2010, 18:54
revolution wrote: JohnFound: Does Fresh make use of the fasm macros like stdcall/resource/etc. from the download? Or do you provide you own custom versions of these macros? I ask because these "uncore" updates to macros might affect Fresh users also. Fresh does not use FASM macroses and depends only on FASM core. But this is completely irrelevant. Actually I don't care if I don't have latest updates. Actually for 5 years I created programs with FASM 1.59.3 and I didn't have any problems - this code is still running 24h a day.
Update, only because of "to be updated" is stupid in my opinion. Personally I prefer a world, where people write software with old FASM version, instead of world where all have the latest minor update of FASM, only to use it as a toy. So people, go write some software for the others! |
|||
|
28 Nov 2010, 18:54 |
|
|
killasmurf86 28 Nov 2010, 21:29
Tomasz Grysztar wrote:
A nice way, to have keep version numbers and continue development is to have public source code repository...
When you feel ready you make tag, from which archives are created... Now you discovered small bug. You fix it, and push to repository.... Old tag is available, but everyone know that you have committed bugfix. Also this allows others to see what was modified, and have a closer clue, what, where and why. I can recommend Mercurial... I recently switched to it from git.... for repo hosting I can suggest hgweb and mercurial server If you're interested here's howto for FreeBSD: http://forums.freebsd.org/showthread.php?t=19103 Also you can have a look, how it looks at http://hg.bsdroot.lv/ http://hg.bsdroot.lv/pub/aldis/playd.sh/ (my most activly developed project [lately]) If you want I can even host repos for you on my hobbyistish server
(just a thought) |
|||
|
28 Nov 2010, 21:29 |
|
|
edfed 28 Nov 2010, 21:50
one major/minor update is from 1.68 to 1.69 about call reg.
before, call reg32 was compiled as call reg16 now, it is ok, but it lead to incompatibility with my old sources where i should replace all call reg32 by call reg16, and on some of these old sources, there are a lot of them... |
|||
|
28 Nov 2010, 21:50 |
|
|
Tomasz Grysztar 28 Nov 2010, 21:55
killasmurf86: I know those technologies, I use SVN extensively in my everyday work. It's just that I prefer to keep fasm with its own simple standards. They served me well over they years and I don't feel a need to use anything more "modern" - quite the opposite. fasm is not maintained in such a way as a multi-developer projects are.
|
|||
|
28 Nov 2010, 21:55 |
|
|
guignol 28 Nov 2010, 22:31
JohnFound wrote: So people, go write some software for the others! So, you haven't written any software for others, have you?.. |
|||
|
28 Nov 2010, 22:31 |
|
|
guignol 28 Nov 2010, 22:41
Tomasz Grysztar wrote: killasmurf86: I know those technologies, I use SVN extensively in my everyday work. It's just that I prefer to keep fasm with its own simple standards. They served me well over they years and I don't feel a need to use anything more "modern" - quite the opposite. fasm is not maintained in such a way as a multi-developer projects are. Anyway, you've said it yourself - you can easily increment your builds with your own hands, in any convenient way you see. It may start like this: 1.69.27.16927, and then you simply separetly maintain (now!) the two increments between the last dot. What's so scary? |
|||
|
28 Nov 2010, 22:41 |
|
|
Tomasz Grysztar 28 Nov 2010, 22:48
And what's wrong with date playing the role of this number?
What would really be useful: to have version numbers for Windows headers and for AsmEdit core, since (as opposed to fasm core and fasmw IDE) they have no such numbers maintained right now. But that's a different story. |
|||
|
28 Nov 2010, 22:48 |
|
|
revolution 29 Nov 2010, 01:44
Tomasz Grysztar wrote: And what's wrong with date playing the role of this number? |
|||
|
29 Nov 2010, 01:44 |
|
| Goto page Previous 1, 2, 3 Next < Last Thread | Next Thread > |
Forum Rules:
|