Support for Linux
Support for Linux
Should hMailServer have Linux support?
A full version working under Linux would ROCK, but...
...if you could get it to properly function under WINE/Crossover, that would at least be a start! I too would love to use this server under Linux, and if it were only a few weeks to port the bad boy over... Hell you could even setup a PayPal tip jar or feature bounty to help cover your time. I know I'd chuck in a few pennies!
Martin has right, MailServer uses a lot of Windows functionality, its like rewriting the server in linux all over again from a blank empty project,
anyway, there is a mail server in linux, also in the style of "all-in-one" called "XMailserver", but it does not support IMAP. you will need dovecot or courier, but it does not use a mysql-backend, i think there is a way to compile postfix so to use mysql-backend.
anyway, there is a mail server in linux, also in the style of "all-in-one" called "XMailserver", but it does not support IMAP. you will need dovecot or courier, but it does not use a mysql-backend, i think there is a way to compile postfix so to use mysql-backend.
Well I have to agree with lycaros...
...the PHP admin scripts would do just fine! Hell, if we users needed anything else, we could help extend the PHP!
it's actually cool that we are talking about this product getting ported to linux as one of the security admins that I work with is considering putting up a mail server for his domain, he asked me what I use, told him about some of the compalints I have about most linux mail server products, and recommended hmailserver to him; he said unfortinately he doesn't have a windows box that he can run it on.....I think it would be interesting to see also how many linux and unix administrators would convert their existing email infrastructure this is as well
- Blue Ninja
- Normal user
- Posts: 238
- Joined: 2005-12-31 00:22
- Contact:
Hmmm... Maybe an option to consider would be eventually updating the hMailServer code from C++ to C#, since I'd expect most of the .NET classes that hMailServer would use are probably implemented on Mono. Then development would not branch off into a Win32 and a Linux version - at least except for possibly a few small bits of code. This would be easier to maintain than two very different versions.
Also, the Administrator program could be updated as well, since there's a separate forum thread about updating it. As I said there, to do so for the sake of doing so is probably not a good enough reason, but moving it to .NET would not only make it portable (using GTK# or what is supported of WinForms in Mono) to Linux, but would end the VB / C++ debate over there
Also, the Administrator program could be updated as well, since there's a separate forum thread about updating it. As I said there, to do so for the sake of doing so is probably not a good enough reason, but moving it to .NET would not only make it portable (using GTK# or what is supported of WinForms in Mono) to Linux, but would end the VB / C++ debate over there
Perhaps the best way to port this to linux would be to get hmailserver working and dependable on linux platform without the admin interface (administer with phpadmin) and then if needed maybe we can then look at porting the interface next? I have not actually used phpadmin to administer the application; does it have all of the functionality that the admin interface? At any rate, Martin I hope that you do finally agree to port this application to linux as well, after all, making the application available to linux platform is only going to increase it's exposure as an enterprise and small office email server solution.
- Blue Ninja
- Normal user
- Posts: 238
- Joined: 2005-12-31 00:22
- Contact:
I haven't looked at the other admin options, like PHP and whatnot, so I don't know - maybe nobody needs the admin app. Certainly much of it could be done by editing the config files and database, though that's a bit more cumbersome.lycaros wrote:Well i am afraid that itsn't easy to be done, if admin was writen on C# or C++ or if there was something like wxwidgets libary that could help, but anyway who actually needs the admin app?
Porting the admin app should be easier than porting the server itself. While moving from VB to .NET may be trickier than moving from C++ to C#, the app itself is not overly complex. I think it's doable!
I think that Sysadmins nerver runs X-Windows systems in the servers, so its pointless to start a GUI project for a "daemon" server-app thats gonna be "seted-up" only once in the background of a machine without a monitor.
But anyway i think its good idea to have a port of hmailserver in linux, if there is gonna be a secure and stable protion of this, i think it will turn to be the default mail server in all distros over the smpts sendmail and postfix, and dovecot,couriel will be useless becase hmail has already imap/pop3 support.
But anyway i think its good idea to have a port of hmailserver in linux, if there is gonna be a secure and stable protion of this, i think it will turn to be the default mail server in all distros over the smpts sendmail and postfix, and dovecot,couriel will be useless becase hmail has already imap/pop3 support.
- Blue Ninja
- Normal user
- Posts: 238
- Joined: 2005-12-31 00:22
- Contact:
Good point. I run everything off my machine, so I hadn't taken into account that most people won'tlycaros wrote:I think that Sysadmins nerver runs X-Windows systems in the servers, so its pointless to start a GUI project for a "daemon" server-app thats gonna be "seted-up" only once in the background of a machine without a monitor.
As long as it can be reasonably well administered in some way, either by config files or a web interface (preferable for setting up accounts and such), then a GUI app wouldn't be needed.
Though, for some of us who DO access the server locally, it would be nice - I've never seen a web app that comes close to the functionality of a native desktop app, and don't expect to for quite a while. But I do agree, it wouldn't be a high priority. Maybe it's a task I could tackle at some point...
I think that most people would agree that one of the best features of hmailserver is it's intuitive, easy to use gui; I agree with you lycaros that most admins don't use xwindows on servers (I don't even install it on my servers), but again as long as the phpadmin piece can do the job that hmailserver's gui does, then great; if it can't, then I think that even the linux port should give admins the choice to use the gui (you don't have to run xwindows on your server in order to use xwindows programs on your server; i.e. export your DISPLAY variable to your desktop pc; I also agree that this mailserver project could very easy become the default mail server in many distros
I have the felling that the admin is VB and the server is C++ or i have a serious problem with my eyes (or even with my mind who knows?) lol , so i dont understand you, if moving VB to .NET may be trickier, doesn't that means its more difficult to port the admin?, but again i think that is not necessary to port anything but the admin app, GTK++ Or GTK# i guess they wont help us a lot, anyway i dont really know a lot of linux but i think martin can tell us about.Porting the admin app should be easier than porting the server itself. While moving from VB to .NET may be trickier than moving from C++ to C#, the app itself is not overly complex. I think it's doable!
- Blue Ninja
- Normal user
- Posts: 238
- Joined: 2005-12-31 00:22
- Contact:
Well I can't speak for your mind, but your eyes seem ok
What I meant was, porting VB to VB.NET is trickier (I believe) than C++ to C#. However, the server application is larger and more complicated than the admin app. Therefore, proting the server will be a whole lot of relatively easy work, while proting the admin app will be a little bit of hard work.
There had been some talk of porting the Admin app to C++, in another forum. When updating it to .NET, the language is - for all practical intents and purposes - irrelevant. So it could be made C#, but it wouldn't be any different than VB.NET. In fact, a few clicks will translate between the two pretty nicely.
But since I know VB and not C#, then if I were to do it, I'd do it in VB.
What I meant was, porting VB to VB.NET is trickier (I believe) than C++ to C#. However, the server application is larger and more complicated than the admin app. Therefore, proting the server will be a whole lot of relatively easy work, while proting the admin app will be a little bit of hard work.
There had been some talk of porting the Admin app to C++, in another forum. When updating it to .NET, the language is - for all practical intents and purposes - irrelevant. So it could be made C#, but it wouldn't be any different than VB.NET. In fact, a few clicks will translate between the two pretty nicely.
But since I know VB and not C#, then if I were to do it, I'd do it in VB.
If porting the admin interface would be less work than porting the server, then I vote for porting them both; Looking at some of the most successful cross platform applications such as mysql and apache for example, I would guess that they are probably so successful because users can choose to run them on their windows or linux boxes, without loosing functionality, stability or ease of use from one platform to the next
- Blue Ninja
- Normal user
- Posts: 238
- Joined: 2005-12-31 00:22
- Contact:
Yep, a known issue - follow these instructions: http://www.hmailserver.com/documentation/?page=buildinglycaros wrote:I Agree, i can't download the source codes anyway, the server returns "Mal-formed reply from origin server. Please try your request again."
- Blue Ninja
- Normal user
- Posts: 238
- Joined: 2005-12-31 00:22
- Contact:
Yep - I'd be lost without the MySQL Administrator and other GUI tools! They make things a lot easier. And I think hMailServer Administrator makes local management much simpler. While the GUI is probably not important for many users, it will be a big plus for others.abliss26 wrote:Looking at some of the most successful cross platform applications such as mysql and apache for example, I would guess that they are probably so successful because users can choose to run them on their windows or linux boxes, without loosing functionality, stability or ease of use from one platform to the next
And since it won't be a "big deal" to port, it seems like a good thing to do, at least eventually...
- Blue Ninja
- Normal user
- Posts: 238
- Joined: 2005-12-31 00:22
- Contact:
Well, *I* agree with you... But as some others pointed out, the Admin program wouldn't be used by a lot of Linux administrators who access the server remotely - so from their perspective, only the server itself is a "neccessity"... Honestly, I don't know what percentage of people would fall into that category, versus administering it locally. I guess that statistic would determine how big a priority the admin utility is... In either case, it wouldn't do much good without the server itself, so that would still be the first thing to work on!martin wrote:Eventually? The administration part feels kind of vital and if hMailServer should be ported, it should be ported immediately...
the mail servers in linux has nothing but config files, in some cases there is and a web-administration, ok lets say that hmail in linux will gonna be a something like a mail daemon "for admins only" lol
PS. Is there anyone who can zip the source and send it to me? please? because i cant install trotoiseCVS.
PS. Is there anyone who can zip the source and send it to me? please? because i cant install trotoiseCVS.
Again, I must reiterate (sp) giving administrators the choice of using a gui OR editing config files to setup and administer hmailserver should be just as high of a priority as running the app itself on linux; I'm really not sure why were debating this (if we are); look at it like this; I highly doubt giving admins the choice to use a gui is going to turn linux administrators away from using hmailserver, while not giving administrators the choice might do just that. If there is an implication that if an administrator prefers or chooses to use a gui versus using only a command line interface then they must not be a "real" linux administrator, I would ask then why is so much time and effort spent in developing such great windows managers such as twm, kde and gnome? The truth is, people like choice; it's that simple; being an administrator or using a computer really comes down to getting something done, and that's it; whether that means using a gui, or vi to edit files doesn't really matter as long as the job gets done; (sorry for rambling on here)
Last edited by abliss26 on 2006-01-21 21:30, edited 1 time in total.
- Blue Ninja
- Normal user
- Posts: 238
- Joined: 2005-12-31 00:22
- Contact:
Ahh, ok - I haven't used the PHPWebAdmin, so I didn't know if it was a viable substitute or not. Or whether it worked in Linux. So, I guess it will be a high priority then!martin wrote:Blue Ninja:
But hMailServer would require an administration interface, even if you want to administer it remotely. And PHPWebAdmin wouldn't work under Linux.
Erase the "real linux administrator" and replace it with "real linux dedicated server administrator" i mean the one admin who owns a "dedicated server" (network term - , a dedicated server cannot be used as a workstation) kdm, and gnome, and the rest, is to turn a linux into a desktop, this is why is so called "desktop environments" you can be a real admin of a linux server/desktop or both or whatsoever is enought if you are able to configure it however you want.if an administrator prefers or chooses to use a gui versus using only a command line interface then they must not be a "real" linux administrator
Just i have the felling that there is no any company that hosts linux Dedicated-servers and provides you with a "GUI Remote admin" (like that which windows XP remote admin has build in) instead of a shell.
- Blue Ninja
- Normal user
- Posts: 238
- Joined: 2005-12-31 00:22
- Contact:
I think the point raised earlier was that many admins will be running a Linux server hosted at some faraway hosting company, where they won't have physical access to it - so telnetting in or using a web interface would be the only viable options. For many of these cases, a GUI would be of no value to them.abliss26 wrote:If there is an implication that if an administrator prefers or chooses to use a gui versus using only a command line interface then they must not be a "real" linux administrator, I would ask then why is so much time and effort spent in developing such great windows managers such as twm, kde and gnome?
But I do agree that for many, the GUI is important, if not "critical". And as Martin pointed out above, the PHPWebAdmin won't do the trick, so that makes it even more vital.
- Blue Ninja
- Normal user
- Posts: 238
- Joined: 2005-12-31 00:22
- Contact:
I was goingto post it half an hour ago, but it's STILL downloading... I think something might actually be screwed up. The whole program takes up about 5MB on my system, but the source is almost 50MB so far, most of all which seems to be some library called "boost", I have no idea what that is.lycaros wrote:PS. I really need someone to zip the source and send it to me? please? because i cant install trotoiseCVS.
I'll post it when - and IF - it finishes. By the looks of it, I must be downloading the entire Novell Forge!
- Blue Ninja
- Normal user
- Posts: 238
- Joined: 2005-12-31 00:22
- Contact:
FINALLY... Here it is, zipped down to 8MB...lycaros wrote:We can use a config file and a little tool to transfer the config in the database, its not a big deal, but somebody has send me the source code
http://www.blueninjasoftware.com/hmailserver.zip
I apologize for the crappy outgoing speed at the moment...
P.S. - I guess 72MB isn't that bad, my sources folder for the stuff I've worked on for my clients is over 2.6GB [/url]
Last edited by Blue Ninja on 2006-01-21 22:45, edited 1 time in total.
- Blue Ninja
- Normal user
- Posts: 238
- Joined: 2005-12-31 00:22
- Contact:
- Blue Ninja
- Normal user
- Posts: 238
- Joined: 2005-12-31 00:22
- Contact:
- Blue Ninja
- Normal user
- Posts: 238
- Joined: 2005-12-31 00:22
- Contact:
Until hmail get to work on linux (i guess after a long period), you still can do something in linux that will smell like "HMail" If you compile postfix with mysql support, and configure the dovecot. so to have a mix of postfix/dovecot you can store the virtual mailboxes and the domains in the database but you will have to use phpMyAdmin when you want to make any changes (add accounts/domains etc) but still, the main configuration for the smtp will be in a config file.
I already did the configure and created the tables so if you want i can send you both the 11 preconfigured files and the .sql but again the mix lacks of some features that hmail have
I already did the configure and created the tables so if you want i can send you both the 11 preconfigured files and the .sql but again the mix lacks of some features that hmail have
- Paulo Meireles
- Normal user
- Posts: 86
- Joined: 2005-04-16 13:09
- Location: Lisbon - Portugal
I guess the whole point of running hMailServer under Linux is to have a completely free platform. That's my goal, too. However, it seems that it's not trivial to port hMailServer to Linux. As an intermediate step, it would be very nice to have it running on reactos (http://www.reactos.org) which, although yet incomplete, is already very promising. The porting effort should be nearly nil, and we would have a free platform to run hMailServer. What do you think?
Paulo
Paulo
I'm not against having a version of hmailserver available for reactos, however in a production enviroment, I would only run hmailserver on a linux or windows box; for my home business, the reactos version wouldn't help me personally, as I only have 2 production boxes, 1 running windows and 1 running centos 4. While I agree that it may not be trivial to port hmailserver to linux, I think that doing so will greatly increase hmailserver's expose and present it as a real alternative to linux mail servers such as postfix, qmail and sendmail....
Before reaching too quickly for the "YES!" vote, try rewriting the question to:
Do you want to pull development effort off of making the Windows version even better in order to create a Linux version?
One thing to consider is that a Windows programming guru is not necessarily a Linux programming guru. There are two main parts to a program - the top level stuff that the user gets to see (includes the application logic), and the underlying stuff that uses system resources to do the business with the outside world. Windows and Linux are a tad different in their dealing with that latter part.
Do you want to pull development effort off of making the Windows version even better in order to create a Linux version?
One thing to consider is that a Windows programming guru is not necessarily a Linux programming guru. There are two main parts to a program - the top level stuff that the user gets to see (includes the application logic), and the underlying stuff that uses system resources to do the business with the outside world. Windows and Linux are a tad different in their dealing with that latter part.
Andrew McKay
results are going to favor no of course
the results are of course going to favor no (I would guess) just because this currently only is a windows application, as such most of the people using the application and visiting the forums are going to be only windows admins; that being said, of course a windows only admin would not vote for this to be ported