mORMot Open Source

AOG, thanks, with your roadmap, with the efforts by you and ab and other contributors, I have the confidence to use mORMot for web servers for Linux.

Thanks for your effort, AOG.

BTW, Several general questions about running mORMot under linux -

1 - Do you use the MVC part of the mORMot framework under linux?
2 - Does your mORMot web server server the Internet users standalone or behind something like nginx?
3 - How do you monitor the mORMot web server problems, such as when it's hang, when there were an error, when the traffic is to heavy, etc?

Thanks again.

TSOHITServer should be AOG's derived class, but not a class included by the mORMot framework...

Great AOG!
A - The Windows version with cross-compiling was also done by fpcup?
B - I assume I can just run cross_linux_armhf_rtti.bat to update my copy of fpc/lazarus to the latest svn snapshot? (On windows I think I can configure cmd.exe to use the ShadowSocks proxy which is more stable and fast (with the help of Final Speed))

C - And here is a good news - result of my further experiments:
On a Ubuntu server 14.04  64bit, after added the i386 architecture following this instruction (http://askubuntu.com/questions/454253/h … ntu-64-bit), the MVCServer example successfully started and was running without error!

Anyways, I also have tried the x86_64 target when cross-compiling, and on the same Ubuntu server it failed to start and output the following error, I'm just posting it here as a feedback:

edwin@mysvr:~/download$ ./MVCServerx64
An unhandled exception occurred at $0000000000610037:
EAccessViolation: Access violation
  $0000000000610037  SHA256EXPANDMESSAGEBLOCKS
  $0000000000613D6B  FINAL,  line 4827 of ../../../SynCrypto.pas
  $0000000000613DD1  FULL,  line 4848 of ../../../SynCrypto.pas
  $0000000000611957  SHA256,  line 1621 of ../../../SynCrypto.pas
  $00000000005AEC26  SETPLAINPASSWORD,  line 194 of MVCModel.pas
  $00000000005AED1B  INITIALIZETABLE,  line 209 of MVCModel.pas
  $000000000059B6BF  CREATEMISSINGTABLES,  line 1063 of ../../mORMotSQLite3.pas
  $0000000000401619  main,  line 43 of MVCServer.dpr

@AOG,

Thanks again! Downloaded the windows version, Lazarus started OK, great!  When cross-compiling the MVC server sample it failed, not sure what's wrong. Log messages.

Compile Project, OS: linux, CPU: x86_64, Target: MVCServer: Exit code 1, Errors: 1, Warnings: 189, Hints: 1603
MVCServer.dpr(4,2) Note: APPTYPE is not supported by the target OS

.....

MVCViewModel.pas(448,21) Hint: Function result variable of a managed type does not seem to be initialized
MVCViewModel.pas(443,5) Note: Local variable "auto" is assigned but never used
C:\fpclazrtticross\fpc\units\x86_64-linux\rtl\cprt0.o: In function `_start':
(.text+0x3c): undefined reference to `__libc_csu_init'
C:\fpclazrtticross\fpc\units\x86_64-linux\rtl\cprt0.o: In function `_start':
(.text+0x43): undefined reference to `__libc_csu_fini'
MVCServer.dpr(62,1) Error: Error while linking

What am I missing? Thanks AOG.

And the failed tests:

1.2. Low level types: 
  - RTTI: 410 assertions passed  1.16ms
  - Url encoding: 200 assertions passed  978us
! Low level types - Encode decode JSON
! Exception EAccessViolation raised with messsage:
!  Access violation
  - Variants: 48 assertions passed  234us
  - Mustache renderer: 146 assertions passed  6.15ms
  - TDocVariant: no assertion  326us
  - BSON: no assertion  133us
  - TSynTableStatement: 206 assertions passed  350us
  - TSynMonitorUsage: 1,202 assertions passed  500us
  Total failed: 5 / 284,795  - Low level types FAILED  4.63s

 2.4. Memory based: 
  - Database direct access: 20,246 assertions passed  259.74ms
  - Virtual table direct access: 12 assertions passed  9.83ms
  - TSQLTableJSON: 144,070 assertions passed  311.32ms
  - TSQLRestClientDB: 402,407 assertions passed  2.71s
  - TRecordVersion: 20,060 assertions passed  270.42ms
  - RTree: 140,000 assertions passed  1.18s
!  - Shard write: 100 / 10,602 FAILED  1.22s
!  - Shard read: 20,000 / 140,001 FAILED  1.35s
!  - Shard read after purge: 2 / 114,003 FAILED  1.22s
  Total failed: 20,102 / 991,401  - Memory based FAILED  8.56s

I'm using your fork of mORMot, and it's Much less errors than the version compiled by the official Lazarus public release.

Well done AOG!

@AOG, Thanks for sharing!

I'm still not clear about that, What problems I might have if I use the the official Lazarus 1.6?  Since currently it already compiles the latest mORMot (Arnaud's version, not tested your version yet).

My concern is that, is the version you kindly provided shares the same configuration data with the official Lazarus release? You know, since I'm new to Lazarus, but when upgrading Delphi it's really a pain to configure the paths, the 3rd party libs and tools, etc, I'm not sure if it's also that case with Lazarus.

AOG,

Re. fpcup is not working, because almost all internet connections from inside China going outbound are NOT stable, because of this: (https://en.wikipedia.org/wiki/Great_Firewall), the svn checkout will not get to the end. It's a real pain in the ass, if you ask me...

On the other hand, if I use TortoiseSVN, I can check out this FPC source (https://github.com/graemeg/freepascal). But I haven't had the time to check how to use it with my Lazarus 1.6 installation, if you have any hints, I'll be appreciated...

Thanks for the info ab.

May 13, 2016 updates:
Installed Lazarus 32bit 1.6 under Linux Mint Xfce 32bit and successfully compiled and executed:
- The self tests.
- 01 - In Memory ORM
- 04 - HTTP Client-Server

Wow! That feels so good Thanks Arnaud and other contributors like AOG!

Other notes:
Running self tests emitted quite a lot of errors (same as when being compiled with XE4).

Now I should look around for a way to update my FPC to the latest trunk since fpcup doesn't work on my netowork, so anybody has any hints would be welcomed!

Done a very small donation

@aroxing,

It'll be helpful to others if you report back your usage experiences later

Hey ab, this is a very vulnerable info to me! Thanks!

Hi ab,

I'm not just trying to argue, but I view it as **exceptional** if you use BatchStartAny() and then BatchDelete(ID: TID)  without providing a TSQLTable class, but because it's not expected to be so.

Hi ab,

If you allow me, I'd like to suggest that this design is error-prone - as you can see in my reported case.

Raising exception is a better way for catching errors, because, quoting from SO user Ned Batchelder:

Raising exceptions is a modern software design for error handle, people says.

You can check this StackoverFlow discussion for further info: http://stackoverflow.com/questions/4670 … error-code

Just my 2 cents

AFAIK It's not possible with the current version.

I guess that can be achieved by using attributes (http://wiki.freepascal.org/Property_attributes), which is also supported by FPC (). But I'm afraid Arnaud will have to abandon support for earlier versions of Delphi and Kylix, but I doubt he will.

Another idea would be allow defining a class method for TSQLRecord, something like TSQLRecord.GetAutoPersistedFieldsInCSVFormat(), where people can return a CSV-formatted string like 'ID, Name, Year', to define which fields will be handled by the framework. But I'm not sure if that will work, ab can tell us

hi ab,

I'm not using external FTS4 tables, but I'll provide feedback if I do in the future.

I don't think the .obj file size is a real concern nowadays, all after all, it'll still be in the "light" level. I just read the FTS5 document:

That doesn't sound very interesting...

Thanks ab. The first link is broken. I went through the second forum post.

In short, if I use TByteDynArray instead of TSQLRawBlob as the type for my binary fields, the binary data will be transferred by BatchAdd() and BatchUpdate(), is this correct? Thanks again!

Thanks, is your client html/js or Delphi? I mostly interested in JS (in browser)

I'm not able to help, but I'm interested in this topic.

Well done guys!

Great! Thanks Arnaud!

@ab,  as a similar question, the last time you asked me to have created a ticket about intercepting all **exceptions** in a single place on the **client** side. Are you still interested in that ticket

http://synopse.info/fossil/tktview/f561 … a8ef19b2a4