Recent Posts

5951

Developer's Corner / Re: source control systems: what's the best?

« Last post by f0dder on January 08, 2008, 08:02 AM »

I don't really mind doing "svn update" or "svn commit" or "svn add" et cetera from a command shell, but where TortoiseSVN comes in super-handy is in it's repo-browser, which is easier to navigate around than a lot of "svn ls <remote-uri>".

Also, when doing commits, I use the TortoiseSVN interface for checking the diffs of modified files, when writing the entry changelog. Easier than manually remembering & keeping track of all changes. Works very well. Oh, and the auto-completion in the changelog entry field is also very nice.

5952

Developer's Corner / Re: source control systems: what's the best?

« Last post by f0dder on January 08, 2008, 07:27 AM »

tortoise is a quality product, but it was a bit too "invasive" for me with its shell extensions.
i use RapidSVN (free): http://rapidsvn.tigris.org/

-mouser (January 07, 2008, 11:17 PM)

If RapidSVN only had as nice an UI as Tortoise, and (especially important to me), as nice a diff viewer, repo-viwer, etc...

As for backups, mouser, you should install http://www.rsnapshot.org/ on your server.

5953

Living Room / Re: is someone stealing my bandwidth?

« Last post by f0dder on January 08, 2008, 07:25 AM »

Do you have a WiFi router used with a wireless network card (probably built-in), or do you have one of those WiMaxx connections? Ie., is the WiFi only for your local LAN, or do you have your *internet access* through WiFi?

The wireless-internet connections I've seen have always been oversold (which is scary considering that not so many people in .dk have adopted it), you can very clearly see the traffic impact when people get home from work...

If you have a "regular" internet connection and just a WiFi router/accesspoint, indeed be sure that WPA-SPK (not just WEP, ugh) protection is enabled. And a restart of the accesspoint can also help, like Stoic Joker said, it's amazing how unstable a lot of WiFi APs are >_<

5954

Best Text Editor / Re: Boxer Text Editor

« Last post by f0dder on January 08, 2008, 07:20 AM »

jgpaiva: of course localized UI strings are useful for non-english people, my point was just that those strings shouldn't go in source files but instead in external resources. Non-ASCII has no place in source files.

Sure, it's slightly more work than embedding strings directly, but it pays off in the end. Especially once that little in-house tool has grown enough to be useful for other people, or your little local company merges/hires abroad... and it teaches you better habits as well.

5955

Living Room / Re: Request: fixes for geshi syntax formatter

« Last post by f0dder on January 08, 2008, 07:16 AM »

Hm.

If I only select part of some code (ie, I make sure that I don't select "above" or "below" the code, as you're almost 100% guaranteed to do if you don't stop selection at second-last char (when selection down) or second char (when selection up), I don't get line numbers, but I get #-signs.

5956

General Software Discussion / Re: Heap memory view

« Last post by f0dder on January 08, 2008, 07:12 AM »

For large systems I usually use Paul Nettle's memory manager in debug builds. I have only used HeapMemView on smaller projects in the rare situation where I'm seeing problems in release builds and not in debug builds, but I agree with you, doing things post-mortem is kind of like locking the stable door after the horse has bolted 

-ChalkTrauma (January 07, 2008, 08:45 PM)

Paul Nettle's memory manager is pretty good for a free library, I've used it a bunch of times and it's really helped me out. (And ugh, paulnettle.com has moved entirely to flash, UGH!).

5957

Developer's Corner / Re: source control systems: what's the best?

« Last post by f0dder on January 07, 2008, 06:47 PM »

Forget about VSS (Visual SourceSafe), too many people have had corruption issues, and afaik Microsoft doesn't even use it themselves.

Forget about CVS, it's old and outdated.

I have no experienec with Perforce, so can't speak about that, but I haven't really heard anything negative about it.

Personally I use subversion, which hasn't let me down, has been around long enough that it's stable & mature enough, and has decent performance (even across the internet). RE VisualSVN: Ankh also offers VS SVN integration, and is free.

I'm interested in BZR, but it's still too young for me to touch it.

5958

General Software Discussion / Re: Heap memory view

« Last post by f0dder on January 07, 2008, 06:44 PM »

You really ought to compile your debug builds of larger systems with a heap allocation debugging library... better than using external tools or doing post-mortem analysis.

5959

Living Room / Re: The Rule of 3 Drives: How to Build your Next PC

« Last post by f0dder on January 07, 2008, 06:42 PM »

The speed would be horrible over NAS. SMB is crud. However, you can easily attach it to an existing piece of network equipment if you want NAS. Almost any existing NAS has USB ports for external drives. And, I assume, if you need a NAS, you already have a network, stick it on a server. If you don't have any of those then grab a new ASUS or D-Link wireless appliance and start hanging stuff off of that. The latter is what I do.

-tinjaw (January 07, 2008, 06:24 PM)

Yeah, SMB/CIFS is pretty horrible, I normally seem to get ~25MB/s on a gigabit LAN, sometimes if I'm lucky I can get 30-35MB/s. FTP'ing from a file that's 100% in filesystem cache with ramdisk as destination gives me ~80MB/s stable, and raw benchmarking gives me ~925mbit/s. So it's safe to say that v1 of SMB/CIFS for high-speed networks... can't wait until samba support CIFSv2 though, even if you'd have to upgrade clients to Vista SP1 to get advantage of it :/

What does drobo run internally? Linux w/kernel software raid?

-f0dder (January 07, 2008, 03:02 AM)

"Drobo OS", with which they've been fairly close-lipped about the details.  At least they were when I looked at the Drobo.

-Liquidmantis (January 07, 2008, 08:53 AM)

Sounds fishy... I find it hard to believe they rolled their own, since most other vendors tend to use a linux kernel.

5960

Best Text Editor / Re: Boxer Text Editor

« Last post by f0dder on January 07, 2008, 06:37 PM »

In my translation work I've been working exclusively in Unicode for two or three years now, so I can't escape it, even though I think it's pretty evil Well, UTF-8 is a tolerable necessary evil, but UTF-16 is, IMtotallyHO, true unmitigated concentrated essence of hell-bound evilness, aka the spawn of satan (and I'm being generouos too!)

-tranglos (January 07, 2008, 11:57 AM)

I more or less agree with that. UTF-16 (or rather, UCS-16) is a decent enough internal format (and using UTF-8 for internal is pretty insane and bad). But for exchange, I certainly prefer UTF-8 instead of the other formats (no BOM hell, endian purgatory, etc.)

And yes, I realize that some compilers/interpreters support UTF8 source code files now, but that's primarily Visual Studio, which is already an IDE. Python does as well though, doesn't it?

-tranglos (January 07, 2008, 11:57 AM)

Programming languages should only accept US-ASCII for variable names etc., and imho shouldn't support any of UTF/UCS even for literal strings. It's plain evil. If you want internationalization, do the proper thing and use external files.

5961

Living Room / Re: Request: fixes for geshi syntax formatter

« Last post by f0dder on January 07, 2008, 06:33 PM »

Use Opera, it does not have that problem

-Lashiec (January 07, 2008, 12:17 PM)

Opera doesn't copy line-numbers?

I wonder if that's a bug or a feature or correct handling of the HTML

5962

General Software Discussion / Re: Dealing with spam

« Last post by f0dder on January 07, 2008, 06:30 PM »

Imho if you set up your own SMTP server (unless you mean serious business), it should only be for incoming mails, not for sending outgoing - use your ISP or "something big & well-known" for sending.

Problem is even with a static IP, unless you pay for what some ISPs in Denmark call a "global IP", you won't always have control of the IP's reverse-DNS... people & software get suspicious when my.serious.biz resolves to 1.3.3.7, but 1.3.3.7 reverse-dns is something like 0x12345678.slnxx7.adsl-dhcp.tele.dk instead of my.serious.biz.

5963

Living Room / Re: What is appropriate content for DonationCoder?

« Last post by f0dder on January 07, 2008, 05:10 AM »

Wooop unlock!

5964

General Software Discussion / Re: Speed Load

« Last post by f0dder on January 07, 2008, 05:08 AM »

Hm, unless copy protection is in play, bind+rebase should never invalidate an exe, unless the guy behind speedload is using his own exe modifying routines and haven't coded them properly.

bind+rebase is a good idea, and imho it should be done by installers, but I must confess that I haven't been able to measure much impact by doing it; if you have measured a speed boost, it's probably because you haven't rebooted after the bind, and thus files are cached etc.

Import binding is pretty fast, by the way - DLL exports are always sorted, and thus binary search can be (and is) used when searching for a matching export for an import from a EXE.

5965

Living Room / Re: The Rule of 3 Drives: How to Build your Next PC

« Last post by f0dder on January 07, 2008, 03:02 AM »

If I might comment on the RAID topic, be very wary of where you place your confidence.  After multiple failures on a variety of "SOHO" type RAID implementations (low/moderate cost cards, motherboard built-in offerings, etc.) I've written them off and use non-realtime disk mirroring.  As an added bonus you get some degree of versioning protection that way.

The problem with the budget RAID solutions is that they won't rebuild.  The failures I've had in RAID1 configs made the mirror drive unreadable.  Since moving to a multilayer asynchronous drive mirror solution my file servers at work have been bulletproof.  And from skimming Google's whitepaper on data protection they use a similar system.

-Liquidmantis (January 06, 2008, 09:21 PM)

I wouldn't use any raid level but MIRROR, for a couple of reasons:

• RAID-5 rebuilds are expensive, and might take an additional disk out when rebuilding. You can still only lose one disk without losing all your data.
• With non-mirror RAID levels, it can be difficult moving from one controller type to another (as in, virtually impossible).
• With mirroring, you can keep running from a single drive, if you want/need to.

for MIRRORing, just about any solution will work (100% software done on the OS, half-assed "hardware" solutions like onboard and most cards, and "real hardware"). If you're going to buy a "hardware raid" card, either get one with cache memory and a batteri-backup unit, or don't bother at all. And since you shouldn't be running the fancier RAID modes, don't bother

Now our enterprise level RAID systems recover well, especially when configured with a hotspare.  The only RAID system I'd consider at home now though is a Drobo.

-Liquidmantis (January 06, 2008, 09:21 PM)

What does drobo run internally? Linux w/kernel software raid?

5966

Living Room / Re: The Rule of 3 Drives: How to Build your Next PC

« Last post by f0dder on January 06, 2008, 08:46 PM »

thanks, i'm going to try that in a few days.

BTW, here's a link to a review of some Continuous Backup utilities.

-Armando (January 06, 2008, 08:42 PM)

That review fails to mention whether the apps simply detect change & copy file, or if they use a filter driver and thus only copy over THE EXACT CHANGES. This might seem a bit anal, but it's pretty damn important if you have data files of several gigabytes where only a few megabytes change.

5967

Living Room / Re: The Rule of 3 Drives: How to Build your Next PC

« Last post by f0dder on January 06, 2008, 08:34 PM »

Yes, MirrorFolder will work for you, and it seems very light on resources. You will want to use what it calls "RAID mode" iirc., which is a silly name for it, since it has noting to do with RAID - it's automagic/instantaneous data replication through use of a filter driver, instead of re-scanning files etc. Should work flawlessly for locally connected drives

5968

Living Room / Re: The Rule of 3 Drives: How to Build your Next PC

« Last post by f0dder on January 06, 2008, 08:22 PM »

Haven't had time to properly test MirrorFolder yet (ie., giving it some really stressful testing), but it seems to be able to do the job well enough. The problem is that if the connection to the server is lost, iirc you have some relatively long timeouts when working with your local files. But I guess there's not much you can do about this, as long as windows filesharing is used, instead of a custom protocol.

5969

General Software Discussion / Re: Janino another new toy?

« Last post by f0dder on January 06, 2008, 08:16 PM »

Small and simple, 1.7meg download... hehe. I'd rather invest some time in http://www.lua.org . But thanks for the link nonetheless

Btw, direct link to janino, I always have trouble finding that stuff on those download sites.

PS: what's the "&promo=100500" part of your link?

5970

Living Room / Re: The Rule of 3 Drives: How to Build your Next PC

« Last post by f0dder on January 06, 2008, 08:13 PM »

(Thanks Lashiec)
f0dder: you use your RAID MIRROR as an "instant restore" backup system ? impossible to use that system with a normal laptop (not talking alienware here...) i guess...

-Armando (January 06, 2008, 08:09 PM)

I use RAID MIRROR as a "oh fsck, the drive died, thankfully I can continue running now (although I should wait until I have a spare disk and can reconstruct mirror), and thankfully I haven't lost up to 23 hours of backups". IN ADDITION TO raid-mirror of your data, you should do regular (and preferably automated) backups to an external storage media (external harddrive enclosure, local or remote fileserver, ...).

No, RAID MIRROR wouldn't really work for a laptop, but neither would mouser's 3 (physical) drive solution

An alternative to RAID MIRROR in your workstation, you could do real-time syncing (and I do mean realtime, not something that runs every hour) to a (local, don't even try realtime sync remotely) fileserver with RAID MIRROR... and then do backups from there too.

Oh, and don't even think about RAID-5 (aka parity).

5971

Living Room / Re: What kind of tagging system would be appropriate for DC?

« Last post by f0dder on January 06, 2008, 08:08 PM »

I realise there's more to the tag-project than just safeguarding sensitive eyeballs but how about linking the subject text color to certain tags - for example the NSFW threads could have a dark maroon subject line which would blend in with the black and yet be distinguishable. Maybe we could have visible tags (noise!) on by default but once one knows the color legend it could be switched off...

-nosh (January 06, 2008, 03:16 PM)

Please, not by default. People wanting the kool-aid can go drink it in their personal preferences section

-f0dder (January 06, 2008, 07:02 PM)

Fine! I was just worried about some poor child strolling by innocently and clicking *God forbid* on an NSFW link. 

-nosh (January 06, 2008, 08:06 PM)

Fortunately we don't have many of those on the forum (and I don't think that would change if a voluntary tagging system as opposed to a forced NSFW system is implemented), and the ones we've had have hardly been anything to lose sleep over (unless you're of the overprotective cotton-cushioning kind). But that's of course my personal POV, having dealt with the Real World^TM from a relatively young age, and not taken supermuch damage from it

5972

General Software Discussion / Re: Dealing with spam

« Last post by f0dder on January 06, 2008, 07:46 PM »

So you are saying that the greylist keeps a list of approved servers (IP addresses?) not domain names? If an email arrives from an unapproved server it is rejected as a temporary error and the server is added if it is resent/received from the same server and the email accepted for processing further?

Sorry I thing I was getting muddled because the terminology was a bit muddled between servers and domains.

If this becomes a widespread tool wouldn't it be easy for spammers to simply check for returned mail and send it out again by beefing up their zombie servers - opening the floodgates to spam and also incidentally validating the email address (because it isn't simply bounced/ignored) so you get spammed even more ?

-Carol Haynes (January 06, 2008, 07:18 PM)

Well, I must admit I don't know 100% how it's implemented internally, and from the tuffmail FAQ it sounds like it's a bit different than what I thought - I was commenting from end-user experience using a greylisting mailserver 

The first time we see an IP address/sender/recipient tripple, and the sender/server meets one of the criteria for Greylisting, the message will be rejected with a temporary error code. A message from an SMTP server that attempts delivery 5 minutes or more after the first delivery attempt to the same IP address/sender/recipient tripple will be accepted.

Yes, I guess that if this was very widespread, spammers might implement more of the SMTP protocol and do proper re-send attempts. But it would probably stop the "dictionary" sends (ie, it tries a crapload of [email protected] and not just web-harvested addresses). Doing things "proper" would make it a lot slower for them to send out their mail, so they'd have to be hard pressed to do it.

The obvious fix is of course to ban all SMTP servers that don't require authentication to send, to be very careful about relaying (if I contact the SMTP server at your.domain and say I have a mail for [email protected] , your.domain should NOT accept the mail¹). But it would require an internet-wide effort to do that, so it's not going to happen.

In general, imho "client machines" shouldn't make SMTP connections to whatever.domain to deliver mail, they should go through a mailprovider with a trusted.recognized.domain and relay mail through there - all other hosts could then be denied for incoming mail. But this would mean a lot of administrative mess in keeping up with who's big and trusted and recognized, and would also require a helluvalot of servers to be reconfigured. Not going to happen.

1: an exception is of course ISPs and other mail providers, but with authentication in place they won't work as open relays, which is what's dangerous.

5973

Living Room / Re: The Rule of 3 Drives: How to Build your Next PC

« Last post by f0dder on January 06, 2008, 07:15 PM »

Renegade: flash-based SSDs made these years will probably outlast harddrives, since they aren't prone to mechanical failures, the amount of erase-cycles have been greatly improved, and there's internal wear-leveling algorithms in use as well. Thing with current SSDs is that while their random seek performance is super++, reads and especially writes are still relatively slow, unless you go for the top-of-the-line and sell your kidneys.

Also note that you should use physical drives, not just partitions on the same drive.

And (somebody please link the old thread ), I would really recommend that you go for one fast system+apps drive (raptor is a good bet), and then chain 2 drives to a RAID MIRROR. Mouser's 3-drive suggestion won't really help you if your data drive crashes, a mirror will (yeah sure, you should be doing automated backups, but if you only do that once per day, you can lose up to 23 hours of work if your data drive crashes. And what if it crashes during a backup? in-con-sis-ten-cy).

Personally I have 2x74gig raptor drives in my box, and a 400gig mirror on a fileserver on a gigabit LAN; too bad windows filesharing is so slow (expect to get 25-30megabyte/sec top, even though the LAN can easily go 90+ meg/s), but it's a lot of peace of mind this way, especially since my bulk storage is separated from my main box.

If I didn't have a fileserver, I'd have gone for one or two raptors, and a raid mirror in my own box. Perhaps external firewire or eSATA case as well, but I don't feel that I need that with my fileserver and all.

And oh, RAID isn't a replacement for backups, it's just a safeguard from when your drive dies from mechanical wear and tear.

5974

General Software Discussion / Re: Dealing with spam

« Last post by f0dder on January 06, 2008, 07:06 PM »

Sorry - I think I am getting confused here - what do you mean by 'domain' in this context.

For me domain means the bit after the @ symbol so if I have some people who write to me from a common domain name such as hotmail.com which are perfectly legitimate emails so they correctly bounce back to me after being rejected then won't everything addressed to hotmail.com addresses also be validated without further checking?

-Carol Haynes (January 06, 2008, 04:21 PM)

validated by the greylisting filter, yes. And only for mail originating from those servers, not spoofed to be from those servers. Spam filtering kicks in after greylisting to get rid of actual spam sent from a greylist-okay domain.

The idea with greylisting is that most spamming is done from very brutish zombie software that doesn't implement much of the SMTP protocol, namely re-transmits on temporary error.

5975

Living Room / Re: What kind of tagging system would be appropriate for DC?

« Last post by f0dder on January 06, 2008, 07:02 PM »

I realise there's more to the tag-project than just safeguarding sensitive eyeballs but how about linking the subject text color to certain tags - for example the NSFW threads could have a dark maroon subject line which would blend in with the black and yet be distinguishable. Maybe we could have visible tags (noise!) on by default but once one knows the color legend it could be switched off...

-nosh (January 06, 2008, 03:16 PM)

Please, not by default. People wanting the kool-aid can go drink it in their personal preferences section