skip to main |
skip to sidebar
Look to get employed today
On Fri, 2010-06-04 at 17:08 +0200, Jason Harris wrote: On Jun 4, 2010, at 4:39 PM, Matt Mackall wrote: On Fri, 2010-06-04 at 14:25 +0200, Jason Harris wrote: Thanks, I of course have debated about this but there are several reasons not to use the standard system hg. What you have here is actually a laundry list of reasons for me to disrecommend and not support people using your tool. Wow! Have you ever even used MacHg? Of course not. I use Mercurial on Linux on the command line pretty much exclusively. I've also never used TortoiseHG, but I have no problem recommending it because they've figured out not to do these things. - Use a current hg. It's a serious disservice to your users to do anything else, as we can't fix their bugs. And we're going to be pretty annoyed when we hear about them again. The point is to isolate the Mercurial version so its stable, while MacHg changes go ahead. See eg Mercurial 1.5.2->1.5.3 if I had used the recent version I keep my users more on the bleeding edge. I don't want to do that. The software that people should use should be long term stable. I am trying for that. And that, as I've said, is misguided. Yes, there was a regression in 1.5.2. But the number of bugs fixed between 1.4.x and 1.5.x is much bigger than that. And when your users encounter bugs, we can only say "we fixed that already, go yell at Jason or stop using MacHG" or "ok, we fixed that, now wait several months for Jason to catch up or stop using MacHG". It's a bad trade-off. Further, I'm not excited about working with add-on developers who _don't care enough about the project_ to run the tip of code. They have an incredibly annoying habit of saying things like "hey, I installed the release you made yesterday and immediately found a regression with my tool I could have told you about months before it shipped had I tested it even once but now it's an urgent problem." But I AM excited to work with devs who ship the latest hg and build nightlies for their users to give us immediate feedback. - Don't add private patches to hg. If you need something patched in hg, send us a patch. Double demerits for using buggy patches to deal with bugs in your tool. Huhh? I have asked about this feature repeatedly on the lists. Best response I got was from Benoit. Basically your response was "Tough luck!' well how was I as a developer of MacHg meant to work around that? I did what I had to do to make the product work. I have been nothing but friendly about asking for a way around this problem.... Sometimes the answer will be "tough luck". Mercurial does what it's doing with temp files in the working directory because it has to. If you can figure out a better way without making life hard for Mac's silly filesystem notifier, that's great, please send a patch. But knowingly breaking Mercurial and shipping it: extremely not cool. - If you run into bugs or performance regressions, tell us about them. Preferably PROMPTLY. Yes, I'm still pissed about that. What? You are pissed? At what? When did you discover the 20% status performance regression? And when did you tell me about it? Maybe I'm mistaken, but I gathered they were not on the same day, as you mentioned it as an excuse for not using 1.5, rather than in a bug report. This says to me "I only care about MacHG". In short, come out of your walled garden and work with the community. What!?! What are you talking about? In what way haven't I worked with the community?!? My source code is open, I am responsive to the issues posted to me... I have posted often to the mailing lists... I am talking about you isolating yourself from the rest of the Mercurial community by shipping a hacked old version of Mercurial rather than supporting the community by using and shipping the latest. (And yes, you've posted often to the mailing list, but almost exclusively and repeatedly about matters that were settled years before you showed up: Q: What's up with error codes? A: Someone who cares[1] should audit those. Q: What about a fixed API? A: No thanks. Q: What about undo in core? A: No thanks. I don't have time to read all the mail I get (never mind reply to it), so extended threads on the above probably aren't going to be a priority for me compared to looking at bug reports and reviewing patches and coding.) Frankly I am kind of stunned at your email here.... I'm glad I've gotten your attention. 1.6 ships on July 1st. Please join us. [1] Which was a hint that YOU should step up and do it, rather than aggravating me into doing it, which is what actually happened. -- Mathematics is the supreme nostalgia of our time. _______________________________________________ Mercurial mailing list Merc ... @selenic.com http://selenic.com/mailman/listinfo/mercurial Hi everybody, The StormOS Hardy Hail Beta 2 is now available for download here< http://stormos.org/node/59 >. StormOS is the first desktop distribution based on Nexenta Core Platform 2 which combines the power of the Solaris kernel with the ease of use of Ubuntu. It aims to be a lightweight OS with everything the average user would want out-of-the-box. The project was started due to my dissatisfaction with OpenSolaris and IPS. Features: - Polished Xfce 4 Desktop enviroment - Abiword word processor - Gnumeric spreadsheet - Gimp - GNU Image Manipulation Program - Rhythmbox Music Player with support for iPod + iTunes sharing and Last.fm - Risretto image viewer - Fast and customizable Firefox 3 web browser - Easy package management with Synaptic Package Manager and gdebi - Transmission bittorrent client - Evince PDF viewer - File-roller archive manager with Thunar integration New in Beta 2: - Updated packages - Faster installer that installs in around 5 mins - Better VM support, now works in 32-bit VirtualBox and VMware - Live CD + Install CD are now part of a single ISO - Bug fixes all 'round Know issues: - Live Session wont work in VMware but it will work fine once installed - Still have GNOME icon in GDM theme I welcome feedback on the forums or IRC channel (#stormos on irc.freenode.net). Regards, Andy (stormos.org) _______________________________________________ opensolaris-announce mailing list open ... @opensolaris.org http://mail.opensolaris.org/mailman/listinfo/opensolaris-announce
